A neurobot (NR) is a mechatronic wearable robot that can be applied to drive a paralyzed limb. Through the application of controllable forces, a NR can assist, replace, or retrain a certain motor function. Robotic intervention in rehabilitation of motor disorders has a potential to improve traditional therapeutic interventions. Because of its flexibility, repeatability and quantifiability, NRs have been more and more applied in neurorehabilitation. Furthermore, combination of NRs with functional electrical stimulation/therapy constitutes a trend to overcome a number of practical limitations to widespread the application of NRs in clinical settings and motor control studies. In this review, we examine the motor learning principles, robotic control approaches and novel developments from studies with NRs and hybrid systems, with a focus on rehabilitation of the lower limbs.