diff --git a/dargs/__init__.py b/dargs/__init__.py
index 41a3f4d..858e9a4 100644
--- a/dargs/__init__.py
+++ b/dargs/__init__.py
@@ -1,3 +1,3 @@
-from .dargs import Argument, Variant
+from .dargs import Argument, Variant, ArgumentEncoder
 
-__all__ = ["Argument", "Variant"]
\ No newline at end of file
+__all__ = ["Argument", "Variant", "ArgumentEncoder"]
\ No newline at end of file
diff --git a/dargs/dargs.py b/dargs/dargs.py
index c67885f..5ea6ad2 100644
--- a/dargs/dargs.py
+++ b/dargs/dargs.py
@@ -23,6 +23,7 @@
 from copy import deepcopy
 from enum import Enum
 import fnmatch, re
+import json
 
 
 INDENT = "    " # doc is indented by four spaces
@@ -632,4 +633,51 @@ def trim_by_pattern(argdict: dict, pattern: str,
                                 f"following reserved names: {', '.join(conflict)}")
     unrequired = list(filter(rem.match, argdict.keys()))
     for key in unrequired:
-        argdict.pop(key)
\ No newline at end of file
+        argdict.pop(key)
+
+
+class ArgumentEncoder(json.JSONEncoder):
+    """Extended JSON Encoder to encode Argument object:
+
+    Examples
+    --------
+    >>> json.dumps(some_arg, cls=ArgumentEncoder)
+    """
+    def default(self, obj) -> Dict[str, Union[str, bool, List]]:
+        """Generate a dict containing argument information, making it ready to be encoded
+        to JSON string.
+
+        Note
+        ----
+        All object in the dict should be JSON serializable.
+
+        Returns
+        -------
+        dict: Dict
+            a dict containing argument information
+        """
+        if isinstance(obj, Argument):
+            return {
+                "object": "Argument",
+                "name": obj.name,
+                "type": obj.dtype,
+                "optional": obj.optional,
+                "alias": obj.alias,
+                "doc": obj.doc,
+                "repeat": obj.repeat,
+                "sub_fields": obj.sub_fields,
+                "sub_variants": obj.sub_variants,
+            }
+        elif isinstance(obj, Variant):
+            return {
+                "object": "Variant",
+                "flag_name": obj.flag_name,
+                "optional": obj.optional,
+                "default_tag": obj.default_tag,
+                "choice_dict": obj.choice_dict,
+                "choice_alias": obj.choice_alias,
+                "doc": obj.doc,
+            }
+        elif isinstance(obj, type):
+            return obj.__name__
+        return json.JSONEncoder.default(self, obj)
diff --git a/tests/test_docgen.py b/tests/test_docgen.py
index 89ba141..e5653d3 100644
--- a/tests/test_docgen.py
+++ b/tests/test_docgen.py
@@ -1,6 +1,7 @@
 from context import dargs
 import unittest
-from dargs import Argument, Variant
+import json
+from dargs import Argument, Variant, ArgumentEncoder
 
 
 class TestDocgen(unittest.TestCase):
@@ -16,6 +17,7 @@ def test_sub_fields(self):
             ], doc="sub doc." * 5)
         ], doc="Base doc. " * 10)
         docstr = ca.gen_doc()
+        jsonstr = json.dumps(ca, cls=ArgumentEncoder)
         # print("\n\n"+docstr)
 
     def test_sub_repeat(self):
@@ -29,6 +31,7 @@ def test_sub_repeat(self):
             ], doc="sub doc." * 5)
         ], doc="Base doc. " * 10, repeat=True)
         docstr = ca.gen_doc()
+        jsonstr = json.dumps(ca, cls=ArgumentEncoder)
         # print("\n\n"+docstr)
 
     def test_sub_variants(self):
@@ -66,6 +69,7 @@ def test_sub_variants(self):
             ], optional=True, default_tag="type1", doc="another vnt")
         ])
         docstr = ca.gen_doc(make_anchor=True)
+        jsonstr = json.dumps(ca, cls=ArgumentEncoder)
         # print("\n\n"+docstr)
 
     def test_multi_variants(self):
@@ -110,6 +114,7 @@ def test_multi_variants(self):
             ])
         ])
         docstr = ca.gen_doc()
+        jsonstr = json.dumps(ca, cls=ArgumentEncoder)
         # print("\n\n"+docstr)
 
     def test_dpmd(self):