Abstract: With the increasing of wind capacity integrated into grid, the traditional deterministic optimization method can hardly meet the requirements for the safe operation of the power system. A multi-objective dynamic dispatch model for power grid with wind farms is presented by considering risk index, which includes such three objectives as minimum fuel cost, minimum emissions and minimum risk index, and also takes the uncertainty of load and the power output of wind farms into consideration. To deal with random variables in this model, probabilistic sequence theory is introduced and its operational space is extended. Then, an improved multi-objective teaching-learning-based optimization (IMOTLBO) algorithm is proposed to solve the model. In the end, the validity and effectiveness of proposed model and algorithm are verified through a 10-gnerators test system with wind farms.