Brain Machine Interface (BMI) is a neuroprosthetic approach for the restoration of motor functions of paralyzed people. While BMI has demonstrated the feasibility of upper-limb neuroprosthesis, it has not yet been evaluated for the restoration of lower-limb motor functions for the disables. In this article we address the following questions 1) whether different step gait related activities can be captured in parallel from rat’s primary motor cortex during walking, and 2) whether this neural information can be used for locomotion or not. We found two categories of cortical neurons: one with their firing rate highly tuned and the other with less or untuned with the step gait cycles. The activities of highly tuned neurons show strong relationship of their firing rates with step gait cycles. Finally, we developed a neuroprosthetic device that stimulates the below of transacted spinal cord from this cortical recording. We propose BMI as a solution for lowerlimb neuroprosthesis.