A fabrication method for a CMOS image sensory device is described. An isolation layer is formed in the substrate to isolate a photodiode sensory region and a transistor device region. A gate structure is further formed on the transistor device region, followed by forming concurrently a source/drain region in the transistor device region beside the side of the gate structure and a doped region in the photodiode sensory region. Thereafter, a self-aligned block is formed on the photodiode sensory region, followed by forming a protective layer on the substrate.