This paper proposed a novel prosthetic hand system integrated with a designed hybrid control method for achieving human-like grasping. The system design has been elaborated, including the mechanism design of a humanoid prosthetic hand equipped with multimodal sensing capabilities, a multi-channel EMG collector, a sensory feedback device, and a real-time central processing unit. A hybrid control method that combines electromyogram (EMG) and visual information was proposed to recognize the control intention of the system user. Preliminary verification of grasping experiments had been conducted, which proved the effectiveness of the prosthetic hand system and the hybrid control method. The development of this system aims to enhance the control performance and perceptual interaction capabilities of prosthetic hands while addressing limitations encountered in traditional myoelectric prosthetic hands. Besides, it could serve as a validation platform for assessing the control performance of diverse algorithms applied in practical applications.