diff --git a/python/pyspark/sql/session.py b/python/pyspark/sql/session.py
index 8f4809907b599..688d67d10d320 100644
--- a/python/pyspark/sql/session.py
+++ b/python/pyspark/sql/session.py
@@ -104,6 +104,46 @@ def toDF(self, schema=None, sampleRatio=None):
     RDD.toDF = toDF  # type: ignore[assignment]
 
 
+# TODO(SPARK-38912): This method can be dropped once support for Python 3.8 is dropped
+# In Python 3.9, the @property decorator has been made compatible with the
+# @classmethod decorator (https://docs.python.org/3.9/library/functions.html#classmethod)
+#
+# @classmethod + @property is also affected by a bug in Python's docstring which was backported
+# to Python 3.9.6 (https://github.com/python/cpython/pull/28838)
+class classproperty(property):
+    """Same as Python's @property decorator, but for class attributes.
+
+    Example:
+
+    >>> class Builder:
+    ...
+    ...    def build(self):
+    ...        return MyClass()
+    ...
+    >>> class MyClass:
+    ...
+    ...     @classproperty
+    ...     def builder(cls):
+    ...         print("instantiating new builder")
+    ...         return Builder()
+    >>> c1 = MyClass.builder
+    instantiating new builder
+    >>> c2 = MyClass.builder
+    instantiating new builder
+    >>> c1 == c2
+    False
+    >>> isinstance(c1.build(), MyClass)
+    True
+    """
+
+    def __get__(self, instance: Any, owner: Any = None) -> "SparkSession.Builder":
+        # The "type: ignore" below silences the following error from mypy:
+        # error: Argument 1 to "classmethod" has incompatible
+        # type "Optional[Callable[[Any], Any]]";
+        # expected "Callable[..., Any]"  [arg-type]
+        return classmethod(self.fget).__get__(None, owner)()  # type: ignore
+
+
 class SparkSession(SparkConversionMixin):
     """The entry point to programming Spark with the Dataset and DataFrame API.
 
@@ -142,8 +182,9 @@ class Builder:
         """Builder for :class:`SparkSession`."""
 
         _lock = RLock()
-        _options: Dict[str, Any] = {}
-        _sc: Optional[SparkContext] = None
+
+        def __init__(self) -> None:
+            self._options: Dict[str, Any] = {}
 
         @overload
         def config(self, *, conf: SparkConf) -> "SparkSession.Builder":
@@ -247,13 +288,19 @@ def getOrCreate(self) -> "SparkSession":
             >>> s1.conf.get("k1") == "v1"
             True
 
+            The configuration of the SparkSession can be changed afterwards
+
+            >>> s1.conf.set("k1", "v1_new")
+            >>> s1.conf.get("k1") == "v1_new"
+            True
+
             In case an existing SparkSession is returned, the config options specified
             in this builder will be applied to the existing SparkSession.
 
             >>> s2 = SparkSession.builder.config("k2", "v2").getOrCreate()
-            >>> s1.conf.get("k1") == s2.conf.get("k1")
+            >>> s1.conf.get("k1") == s2.conf.get("k1") == "v1_new"
             True
-            >>> s1.conf.get("k2") == s2.conf.get("k2")
+            >>> s1.conf.get("k2") == s2.conf.get("k2") == "v2"
             True
             """
             with self._lock:
@@ -276,8 +323,18 @@ def getOrCreate(self) -> "SparkSession":
                     ).applyModifiableSettings(session._jsparkSession, self._options)
                 return session
 
-    builder = Builder()
-    """A class attribute having a :class:`Builder` to construct :class:`SparkSession` instances."""
+    # TODO(SPARK-38912): Replace @classproperty with @classmethod + @property once support for
+    # Python 3.8 is dropped.
+    #
+    # In Python 3.9, the @property decorator has been made compatible with the
+    # @classmethod decorator (https://docs.python.org/3.9/library/functions.html#classmethod)
+    #
+    # @classmethod + @property is also affected by a bug in Python's docstring which was backported
+    # to Python 3.9.6 (https://github.com/python/cpython/pull/28838)
+    @classproperty
+    def builder(cls) -> Builder:
+        """Creates a :class:`Builder` for constructing a :class:`SparkSession`."""
+        return cls.Builder()
 
     _instantiatedSession: ClassVar[Optional["SparkSession"]] = None
     _activeSession: ClassVar[Optional["SparkSession"]] = None