diff --git a/src/arclet/alconna/_internal/_analyser.py b/src/arclet/alconna/_internal/_analyser.py
index 174f4bcd..94d56a9d 100644
--- a/src/arclet/alconna/_internal/_analyser.py
+++ b/src/arclet/alconna/_internal/_analyser.py
@@ -185,7 +185,7 @@ def process(self, argv: Argv[TDC]) -> Self:
         sub = argv.context = self.command
         name, _ = argv.next(sub.separators)
         if name != sub.name:  # 先匹配节点名称
-            if argv.fuzzy_match and levenshtein(name, sub.name) >= config.fuzzy_threshold:
+            if argv.fuzzy_match and levenshtein(name, sub.name) >= argv.fuzzy_threshold:
                 raise FuzzyMatchSuccess(lang.require("fuzzy", "matched").format(source=sub.name, target=name))
             raise ParamsUnmatched(lang.require("subcommand", "name_error").format(target=name, source=sub.name))
 
diff --git a/src/arclet/alconna/_internal/_argv.py b/src/arclet/alconna/_internal/_argv.py
index 01891cb2..115d6948 100644
--- a/src/arclet/alconna/_internal/_argv.py
+++ b/src/arclet/alconna/_internal/_argv.py
@@ -20,6 +20,8 @@ class Argv(Generic[TDC]):
     namespace: Namespace = field(default=config.default_namespace)
     fuzzy_match: bool = field(default=False)
     """当前命令是否模糊匹配"""
+    fuzzy_threshold: float = field(default=0.6)
+    """模糊匹配阈值"""
     preprocessors: dict[type, Callable[..., Any]] = field(default_factory=dict)
     """命令元素的预处理器"""
     to_text: Callable[[Any], str | None] = field(default=lambda x: x if isinstance(x, str) else None)
diff --git a/src/arclet/alconna/_internal/_handlers.py b/src/arclet/alconna/_internal/_handlers.py
index da96a64b..fbfca465 100644
--- a/src/arclet/alconna/_internal/_handlers.py
+++ b/src/arclet/alconna/_internal/_handlers.py
@@ -154,7 +154,7 @@ def step_keyword(argv: Argv, args: Args, result: dict[str, Any]):
                 if arg.value.base.exec(may_arg).flag == "valid":  # type: ignore
                     raise InvalidParam(lang.require("args", "key_missing").format(target=may_arg, key=arg.name))
             for name in args.argument.keyword_only:
-                if levenshtein(_key, name) >= config.fuzzy_threshold:
+                if levenshtein(_key, name) >= argv.fuzzy_threshold:
                     raise FuzzyMatchSuccess(lang.require("fuzzy", "matched").format(source=name, target=_key))
             raise InvalidParam(lang.require("args", "key_not_found").format(name=_key))
         arg = args.argument.keyword_only[_key]
@@ -261,7 +261,7 @@ def handle_option(argv: Argv, opt: Option) -> tuple[str, OptionResult]:
     elif name in opt.aliases:
         error = False
     if error:
-        if argv.fuzzy_match and levenshtein(name, opt.name) >= config.fuzzy_threshold:
+        if argv.fuzzy_match and levenshtein(name, opt.name) >= argv.fuzzy_threshold:
             raise FuzzyMatchSuccess(lang.require("fuzzy", "matched").format(source=opt.name, target=name))
         raise InvalidParam(lang.require("option", "name_error").format(source=opt.name, target=name))
     name = opt.dest
@@ -477,17 +477,17 @@ def analyse_header(header: Header, argv: Argv) -> HeadResult:
     if _str:
         argv.rollback(may_cmd)
         if argv.fuzzy_match:
-            _handle_fuzzy(header, head_text)
+            _handle_fuzzy(header, head_text, argv.fuzzy_threshold)
         raise InvalidParam(lang.require("header", "error").format(target=head_text), head_text)
     if _m_str and may_cmd:
         if argv.fuzzy_match:
-            _handle_fuzzy(header, f"{head_text} {may_cmd}")
+            _handle_fuzzy(header, f"{head_text} {may_cmd}", argv.fuzzy_threshold)
         raise InvalidParam(lang.require("header", "error").format(target=may_cmd), may_cmd)
     argv.rollback(may_cmd)
     raise InvalidParam(lang.require("header", "error").format(target=head_text), None)
 
 
-def _handle_fuzzy(header: Header, source: str):
+def _handle_fuzzy(header: Header, source: str, threshold: float):
     command = header.origin[0]
     if not header.origin[1]:
         headers_text = [str(command)]
@@ -501,7 +501,7 @@ def _handle_fuzzy(header: Header, source: str):
             else:
                 headers_text.append(f"{prefix} {command}")
     for ht in headers_text:
-        if levenshtein(source, ht) >= config.fuzzy_threshold:
+        if levenshtein(source, ht) >= threshold:
             raise FuzzyMatchSuccess(lang.require("fuzzy", "matched").format(target=source, source=ht))
 
 
diff --git a/src/arclet/alconna/config.py b/src/arclet/alconna/config.py
index a1c99590..ac033321 100644
--- a/src/arclet/alconna/config.py
+++ b/src/arclet/alconna/config.py
@@ -109,8 +109,6 @@ class _AlconnaConfig:
 
     command_max_count: int = 200
     """最大命令数量"""
-    fuzzy_threshold: float = 0.6
-    """模糊匹配阈值"""
     _default_namespace = "Alconna"
     """默认命名空间名称"""
     remainders: set[str] = {"--"}
diff --git a/src/arclet/alconna/manager.py b/src/arclet/alconna/manager.py
index db550169..5e5b0907 100644
--- a/src/arclet/alconna/manager.py
+++ b/src/arclet/alconna/manager.py
@@ -116,6 +116,7 @@ def register(self, command: Alconna) -> None:
         self.__argv[command] = __argv_type__.get()(
             command.namespace_config,  # type: ignore
             fuzzy_match=command.meta.fuzzy_match,  # type: ignore
+            fuzzy_threshold=command.meta.fuzzy_threshold,  # type: ignore
             to_text=command.namespace_config.to_text,  # type: ignore
             converter=command.namespace_config.converter,  # type: ignore
             separators=command.separators,  # type: ignore
diff --git a/src/arclet/alconna/typing.py b/src/arclet/alconna/typing.py
index 20598d98..046001d5 100644
--- a/src/arclet/alconna/typing.py
+++ b/src/arclet/alconna/typing.py
@@ -79,6 +79,8 @@ class CommandMeta:
     "命令的作者"
     fuzzy_match: bool = field(default=False)
     "命令是否开启模糊匹配"
+    fuzzy_threshold: float = field(default=0.6)
+    """模糊匹配阈值"""
     raise_exception: bool = field(default=False)
     "命令是否抛出异常"
     hide: bool = field(default=False)