diff --git a/client/starwhale/api/_impl/dataset/builder/mapping_builder.py b/client/starwhale/api/_impl/dataset/builder/mapping_builder.py
index 8f5a672fff..470b1e97be 100644
--- a/client/starwhale/api/_impl/dataset/builder/mapping_builder.py
+++ b/client/starwhale/api/_impl/dataset/builder/mapping_builder.py
@@ -15,13 +15,16 @@
 from loguru import logger
 
 from starwhale.consts import STANDALONE_INSTANCE
+from starwhale.base.uri import URI
 from starwhale.utils.fs import (
     empty_dir,
     ensure_dir,
     blake2b_file,
     BLAKE2B_SIGNATURE_ALGO,
 )
-from starwhale.utils.error import NoSupportError
+from starwhale.base.type import URIType
+from starwhale.base.cloud import CloudRequestMixed
+from starwhale.utils.retry import http_retry
 from starwhale.core.dataset.type import (
     Link,
     BaseArtifact,
@@ -311,11 +314,19 @@ def _handle_bin_sync(self, bin_path: Path) -> None:
             uri, _ = DatasetStorage.save_data_file(bin_path, remove_src=True)
         else:
             sign_name = blake2b_file(bin_path)
-            # TODO: do upload to cloud with new dataset bin storage api
-            uri = sign_name
-            raise NoSupportError(
-                "no support upload bin files into cloud instance directly"
-            )
+            crm = CloudRequestMixed()
+            instance_uri = URI(self.instance_name, expected_type=URIType.INSTANCE)
+
+            @http_retry
+            def _upload() -> str:
+                r = crm.do_multipart_upload_file(
+                    url_path=f"/project/{self.project_name}/dataset/{self.dataset_name}/hashedBlob/{sign_name}",
+                    file_path=bin_path,
+                    instance_uri=instance_uri,
+                )
+                return r.json()["data"]  # type: ignore
+
+            uri = _upload()
 
         self._signed_bins_meta.append(
             MappingDatasetBuilder._SignedBinMeta(
diff --git a/client/tests/sdk/test_dataset.py b/client/tests/sdk/test_dataset.py
index ee4234e9d0..0f80af9cbf 100644
--- a/client/tests/sdk/test_dataset.py
+++ b/client/tests/sdk/test_dataset.py
@@ -1,5 +1,6 @@
 import io
 import os
+import re
 import copy
 import json
 import math
@@ -2130,19 +2131,27 @@ def test_close_empty(self) -> None:
             project_name=self.project_name,
         ).close()
 
-    @pytest.mark.filterwarnings("ignore::pytest.PytestUnhandledThreadExceptionWarning")
     @Mocker()
     def test_put_for_cloud(self, rm: Mocker) -> None:
         instance_uri = "http://1.1.1.1"
         rm.request(
             HTTPMethod.GET,
-            f"{instance_uri}/api/v1/project/self",
+            f"{instance_uri}/api/v1/project/{self.project_name}",
             json={"data": {"id": 1, "name": "project"}},
         )
         update_req = rm.request(
             HTTPMethod.POST, f"{instance_uri}/api/v1/datastore/updateTable"
         )
 
+        server_return_uri = "__server-uri-path__"
+        upload_req = rm.register_uri(
+            HTTPMethod.POST,
+            re.compile(
+                f"{instance_uri}/api/v1/project/{self.project_name}/dataset/{self.dataset_name}/hashedBlob/",
+            ),
+            json={"data": server_return_uri},
+        )
+
         os.environ[SWEnv.instance_token] = "1234"
 
         mdb = MappingDatasetBuilder(
@@ -2164,11 +2173,16 @@ def test_put_for_cloud(self, rm: Mocker) -> None:
         mdb.put(DataRow(index=1, features={"bin": Binary(b"abc")}))
         mdb.flush()
 
-        with self.assertRaisesRegex(
-            threading.ThreadError,
-            "no support upload bin files into cloud instance directly",
-        ):
-            mdb.put(DataRow(index=1, features={"label": 1}))
-            mdb.close()
+        mdb.put(DataRow(index=1, features={"label": 1}))
+        mdb.close()
 
         assert update_req.called
+        assert upload_req.call_count == 1
+
+        assert mdb._signed_bins_meta[0].name == server_return_uri
+        assert any(
+            [
+                server_return_uri in history.text
+                for history in update_req.request_history
+            ]
+        )
diff --git a/client/tests/sdk/test_dataset_sdk.py b/client/tests/sdk/test_dataset_sdk.py
index bbac30d4b2..22dd4d1dd9 100644
--- a/client/tests/sdk/test_dataset_sdk.py
+++ b/client/tests/sdk/test_dataset_sdk.py
@@ -2,6 +2,7 @@
 
 import io
 import os
+import re
 import sys
 import typing as t
 from http import HTTPStatus
@@ -656,6 +657,18 @@ def test_create_for_cloud(self, rm: Mocker) -> None:
             json={"data": {"uploadId": "123"}},
         )
 
+        rm.register_uri(
+            HTTPMethod.HEAD,
+            re.compile("http://1.1.1.1/api/v1/project/self/dataset/mnist/hashedBlob/"),
+            status_code=HTTPStatus.NOT_FOUND,
+        )
+
+        rm.register_uri(
+            HTTPMethod.POST,
+            re.compile("http://1.1.1.1/api/v1/project/self/dataset/mnist/hashedBlob/"),
+            json={"data": "uri"},
+        )
+
         cnt = 10
         for i in range(0, cnt):
             ds.append(
@@ -678,10 +691,7 @@ def test_create_for_cloud(self, rm: Mocker) -> None:
         assert file_request.call_count == 2
 
         # TODO: when sdk supports to upload blobs into cloud, remove assertRasise
-        with self.assertRaisesRegex(
-            RuntimeError, "no support upload bin files into cloud instance directly"
-        ):
-            ds.close()
+        ds.close()
 
     @pytest.mark.skip(
         "enable this test when datastore wrapper supports timestamp version"