The invention belongs to the control field and relates to an intelligent artificial limb movement system based on idiodynamics and control methods thereof. The problems of single recognition mode and low universality of an existing artificial limb system in the prior art are solved. The system comprises an electroencephalogram acquisition module, an electroencephalogram analysis module, a drive control module and a joint angle feedback module. The control methods include the control method in a training mode and the control method in a using mode. The training mode and the using mode are switched through a key. In the training mode, other electroencephalograms are selectively entered as required except a quiescent state, and the quiescent state is a necessary enter item to ensure accurate control of the system. When a user completes electroencephalogram entering in the training mode, a system state is switched to the using mode, and the user can send corresponding electroencephalograms to control the arms to execute corresponding actions.