A bonding structure with a buffer layer, and a method of forming the same are provided. The bonding structure comprises a first substrate with metal pads thereon, a protection layer covered on the surface of the substrate, a first adhesive metal layer formed on the metal pads, a buffer layer coated on the protection layer and the metal pads, a first metal layer covered on the buffer layer, and a second substrate with electrodes and a bonding layer thereon. The first metal layer, the electrodes and the bonding layer are bonded to form the bonding structure. Direct bonding can be performed through surface activation or heat pressure. The method uses fewer steps and is more reliable. The temperature required for bonding the structure is lower. The bonding density between the contacted surfaces is increased to a fine pitch. The quality at the bonding points is increased because fewer contaminations between the contacted surfaces are generated.