Add correlation criteria to the subquery so that a scalar value is returned for each SalesForm row with the corresponding count. You don't need to repeat the SalesForm in the subquery since it's correlated with the outer reference.

UPDATE dbo.SalesForm
SET Pictures = (SELECT COUNT(*)
                FROM dbo.storeImages b
                 WHERE SalesForm.Id = b.WorkOrderId
                );

Your subquery always returns a list of numbers, because you have no condition which matches the entry of SalesForm to the corresponding counted number in your Subquery. Therefore, I added a where clause which matches the entry to update with the entry in the subquery.

UPDATE main
SET Pictures = (SELECT COUNT(b.WorkOrderId) FROM dbo.storeImages b WHERE main.Id = b.WorkOrderId)
FROM dbo.SalesForm AS main

You shouldn't store the count of an existing table, you would need to constantly keep it in sync using triggers or similar.

Instead just use an indexed view. The server will keep it in sync.

CREATE VIEW dbo.StoreImagesById
WITH SCHEMABINDING     -- must be schema-bound
SELECT
  si.WorkOrderId,
  COUNT_BIG(*) AS CountOfImages    -- must use COUNT_BIG not COUNT
FROM dbo.StoreImages si
GROUP BY
  si.WorkOrderId;

CREATE UNIQUE CLUSTERED INDEX pk_WorkOrderId ON dbo.StoreImagesById (WorkOrderId);