diff --git a/dask_bigquery/core.py b/dask_bigquery/core.py
index 5b39b53..2eab9de 100644
--- a/dask_bigquery/core.py
+++ b/dask_bigquery/core.py
@@ -22,7 +22,7 @@
 
 
 @contextmanager
-def bigquery_clients(project_id):
+def bigquery_clients(project_id, credentials: dict = None):
     """This context manager is a temporary solution until there is an
     upstream solution to handle this.
     See googleapis/google-cloud-python#9457
@@ -35,7 +35,15 @@ def bigquery_clients(project_id):
         user_agent=f"dask-bigquery/{dask_bigquery.__version__}"
     )
 
-    with bigquery.Client(project_id, client_info=bq_client_info) as bq_client:
+    # Google library client needs an instance of google.auth.credentials.Credentials
+    if isinstance(credentials, dict):
+        credentials = service_account.Credentials.from_service_account_info(
+            info=credentials
+        )
+
+    with bigquery.Client(
+        project_id, credentials=credentials, client_info=bq_client_info
+    ) as bq_client:
         bq_storage_client = bigquery_storage.BigQueryReadClient(
             credentials=bq_client._credentials,
             client_info=bqstorage_client_info,
@@ -88,6 +96,7 @@ def bigquery_read(
     project_id: str,
     read_kwargs: dict,
     arrow_options: dict,
+    credentials: dict = None,
 ) -> pd.DataFrame:
     """Read a single batch of rows via BQ Storage API, in Arrow binary format.
 
@@ -108,7 +117,7 @@ def bigquery_read(
       NOTE: Please set if reading from Storage API without any `row_restriction`.
             https://cloud.google.com/bigquery/docs/reference/storage/rpc/google.cloud.bigquery.storage.v1beta1#stream
     """
-    with bigquery_clients(project_id) as (_, bqs_client):
+    with bigquery_clients(project_id, credentials=credentials) as (_, bqs_client):
         session = bqs_client.create_read_session(make_create_read_session_request())
         schema = pyarrow.ipc.read_schema(
             pyarrow.py_buffer(session.arrow_schema.serialized_schema)
@@ -132,6 +141,7 @@ def read_gbq(
     max_stream_count: int = 0,
     read_kwargs: dict = None,
     arrow_options: dict = None,
+    credentials: dict = None,
 ):
     """Read table as dask dataframe using BigQuery Storage API via Arrow format.
     Partitions will be approximately balanced according to BigQuery stream allocation logic.
@@ -157,6 +167,9 @@ def read_gbq(
         kwargs to pass to record_batch.to_pandas() when converting from pyarrow to pandas. See
         https://arrow.apache.org/docs/python/generated/pyarrow.RecordBatch.html#pyarrow.RecordBatch.to_pandas
         for possible values
+    credentials : dict, optional
+      Credentials for accessing Google APIs. Use this parameter to override
+      default credentials. The dict should contain service account credentials in JSON format.
 
     Returns
     -------
@@ -164,7 +177,10 @@ def read_gbq(
     """
     read_kwargs = read_kwargs or {}
     arrow_options = arrow_options or {}
-    with bigquery_clients(project_id) as (bq_client, bqs_client):
+    with bigquery_clients(project_id, credentials=credentials) as (
+        bq_client,
+        bqs_client,
+    ):
         table_ref = bq_client.get_table(f"{dataset_id}.{table_id}")
         if table_ref.table_type == "VIEW":
             raise TypeError("Table type VIEW not supported")
@@ -209,6 +225,7 @@ def make_create_read_session_request():
                 project_id=project_id,
                 read_kwargs=read_kwargs,
                 arrow_options=arrow_options,
+                credentials=credentials,
             ),
             label=label,
         )
diff --git a/dask_bigquery/tests/test_core.py b/dask_bigquery/tests/test_core.py
index 67eec62..83b9416 100644
--- a/dask_bigquery/tests/test_core.py
+++ b/dask_bigquery/tests/test_core.py
@@ -336,6 +336,35 @@ def test_read_columns(df, table, client):
     assert list(ddf.columns) == columns
 
 
+@pytest.mark.parametrize("dataset_fixture", ["write_dataset", "write_existing_dataset"])
+def test_read_gbq_credentials(df, dataset_fixture, request, monkeypatch):
+    dataset = request.getfixturevalue(dataset_fixture)
+    credentials, project_id, dataset_id, table_id = dataset
+    ddf = dd.from_pandas(df, npartitions=2)
+
+    monkeypatch.delenv("GOOGLE_DEFAULT_CREDENTIALS", raising=False)
+    # with explicit credentials
+    result = to_gbq(
+        ddf,
+        project_id=project_id,
+        dataset_id=dataset_id,
+        table_id=table_id or "table_to_write",
+        credentials=credentials,
+    )
+    assert result.state == "DONE"
+
+    # with explicit credentials
+    ddf = read_gbq(
+        project_id=project_id,
+        dataset_id=dataset_id,
+        table_id=table_id or "table_to_write",
+        credentials=credentials,
+    )
+
+    assert list(ddf.columns) == ["name", "number", "timestamp", "idx"]
+    assert assert_eq(ddf.set_index("idx"), df.set_index("idx"))
+
+
 def test_max_streams(df, table, client):
     project_id, dataset_id, table_id = table
     ddf = read_gbq(