Spinal cord injury (SCI) induces haemodynamic instability that threatens survival(1-3), impairs neurological recovery(4,5), increases the risk of cardiovascular disease(6,7), and reduces quality of life(8,9). Haemodynamic instability in this context is due to the interruption of supraspinal efferent commands to sympathetic circuits located in the spinal cord(10), which prevents the natural baroreflex from controlling these circuits to adjust peripheral vascular resistance. Epidural electrical stimulation (EES) of the spinal cord has been shown to compensate for interrupted supraspinal commands to motor circuits below the injury(11), and restored walking after paralysis(12). Here, we leveraged these concepts to develop EES protocols that restored haemodynamic stability after SCI. We established a preclinical model that enabled us to dissect the topology and dynamics of the sympathetic circuits, and to understand how EES can engage these circuits. We incorporated these spatial and temporal features into stimulation protocols to conceive a clinical-grade biomimetic haemodynamic regulator that operates in a closed loop. This 'neuroprosthetic baroreflex' controlled haemodynamics for extended periods of time in rodents, non-human primates and humans, after both acute and chronic SCI. We will now conduct clinical trials to turn the neuroprosthetic baroreflex into a commonly available therapy for people with SCI.