Brain-machine interfaces hold promise for restoring basic functions such as movement or speech for severely disabled patients, as well as for controlling neuroprosthetic devices for amputees. One of the major challenges of clinically viable neuroprostheses for chronic use is the implementation of fully implantable recording devices for stable, long-term recordings over large populations of neurons. Recent progress in the microelectronics and MEMS technologies has enabled miniaturization of microelectrode arrays which can be used for recording the neural activity of the brain and for stimulating the neurons. Nevertheless, the demanding requirements of fully implantable recording devices necessitate the incorporation of other active components with the microelectrodes. The presence of the active components in the implant requires power to be delivered to the device. Transcutaneous wires are commonly used for this purpose as well as for communication with the outside world. However, these wires pose the risk of infection to the patient. Another common method of powering is to utilize batteries in the implant. Nevertheless, this method is not desirable as the batteries have to be replaced at the end of their lifetime with a surgical procedure. This thesis presents the development of a remotely powered wireless cortical neural interface system for brain-machine interface applications. In the scope of this work, firstly, the requirements and specifications of the system are analyzed to examine the limits of operation. In addition, the thermal impact of cortical implant operation in the head is investigated to determine the maximum allowable power consumption in the implant. In order to supply power to the active components in the cortical implant, a closed-loop inductive remote powering link operating from a single external battery is implemented. The closed-loop operation enables adaptive delivery of power depending on the implant activity. For improving the power efficiency of the link, a discrete optimization algorithm is developed to design the geometry of the power transmission coils. In addition, application specific integrated circuits are designed to enhance the overall performance of the system and to decrease the footprint of the implant electronics. Moreover, wireless data communication for cortical implants is discussed and a solution for coexistence of the power and data links is presented. For implanting the cortical recording device into the body, a two-body packaging topology is proposed to position the implant inside the cranial bone. With this topology, the performance of the cortical neural interface is significantly improved. The fabricated cortical implants with the two-body package are characterized in air and in vitro. The power transfer efficiency of the closed-loop remote powering link operating from a single supply is measured to be 10.6% in vitro for 10 mW power delivered to the implant load. Finally, the operation of the implant in vitro is validated for over five weeks.