ScreenToWorldPoint has nothing to do with actual scene geometry - it just takes a screen-space + depth vector and converts that point to world-space. I'm assuming you are using ScreenPointToRay in combination with a raycast, in which case you can supply a layermask as a parameter to the raycast and ignore your player's layer;

 //Set this to your player's layer
 [SerializeField] private LayerMask m_LayerMask;
 
 private Camera m_MainCamera;
 
 private void Update ()
 {
     if (m_MainCamera == null)
         m_MainCamera = Camera.main;
 
     //Get the world-space ray from mouse position
     Ray mouseRay = m_MainCamera.ScreenPointToRay (Input.mousePosition);
     //Perform the raycast using the inverted layermask (ignores the selected layers)
     if (Physics.Raycast (mouseRay, out RaycastHit hit, Mathf.Infinity, ~m_LayerMask)
     {
         //Just set this transform to look at the hit point
         transform.rotation = Quaternion.LookRotation (hit.point - transform.position);
     }
 }