This paper addresses the constrained receding horizon control scheme for a class of discrete-time Markovian jump linear systems with partly unknown transition probabilities and persistent bounded disturbances. The proposed method adopts paradigm that utilizes nominal closed-loop feedback controller plus an additional online optimized variable. Both the state feedback approach and the observer based approach are given. The obtained predictive controller guarantees constraint fulfillment and asymptotic state regulation in the mean square sense for all admissible disturbances. The system state and estimated state converge into a stochastic disturbance invariant set provided that the initial state and initial estimation error are feasible. A numerical example is given to verify the effectiveness of the proposed method.
