diff --git a/src/pydantic_to_pyarrow/schema.py b/src/pydantic_to_pyarrow/schema.py
index c52f7a2..ba80ec3 100644
--- a/src/pydantic_to_pyarrow/schema.py
+++ b/src/pydantic_to_pyarrow/schema.py
@@ -111,10 +111,21 @@ def _get_annotated_type(
     return _get_pyarrow_type(field_type, metadata, allow_losing_tz)
 
 
+def _get_dict_type(
+    field_type: Type[Any], metadata: List[Any], allow_losing_tz: bool
+) -> pa.DataType:
+    key_type, value_type = get_args(field_type)
+    return pa.map_(
+        _get_pyarrow_type(key_type, metadata, allow_losing_tz=allow_losing_tz),
+        _get_pyarrow_type(value_type, metadata, allow_losing_tz=allow_losing_tz),
+    )
+
+
 FIELD_TYPES = {
     Literal: _get_literal_type,
     list: _get_list_type,
     Annotated: _get_annotated_type,
+    dict: _get_dict_type,
 }
 
 
diff --git a/tests/test_schema.py b/tests/test_schema.py
index 231a41c..d287589 100644
--- a/tests/test_schema.py
+++ b/tests/test_schema.py
@@ -530,3 +530,28 @@ class EnumModel(BaseModel):
 
     with pytest.raises(SchemaCreationError):
         get_pyarrow_schema(EnumModel)
+
+
+def test_dict() -> None:
+    class DictModel(BaseModel):
+        foo: Dict[str, int]
+
+    expected = pa.schema(
+        [
+            pa.field("foo", pa.map_(pa.string(), pa.int64()), nullable=False),
+        ]
+    )
+
+    objs = [
+        {"foo": {"a": 1, "b": 2}},
+        {"foo": {"c": 3, "d": 4, "e": 5}},
+    ]
+
+    actual = get_pyarrow_schema(DictModel)
+    assert actual == expected
+
+    new_schema, new_objs = _write_pq_and_read(objs, expected)
+    assert new_schema == expected
+
+    # pyarrow converts to tuples, need to convert back to dicts
+    assert objs == [{"foo": dict(t["foo"])} for t in new_objs]