When the nerve cell of human is damaged, the central neural system (CNS) can not work properly. Instead of sending commands by CNS, we can use a micro-stimulation method to send commands to hands, legs, or organs. The first part of this thesis presents a fully implantable system for neural micro-stimulation and neural signal monitoring, and introduces the communication protocol and baseband circuitry of the system. Due to the rapidly development of small and portable electrical equipments, low power becomes more and more important because of the limitation of the battery capacity. Meanwhile, the embedded memory in these devices consumes considerable power. In this thesis, we present a low-power sequential memory decoder to resolve the power-dissipation of embedded memories. We¡¦ll verify that the sequential decoder can reduce the power consumption compared to traditional decoders by implementing our idea with a 2-Kbit SRAM memory.