Abstract: Multi-objective dynamic voltage/reactive power control is a complicated non-linear problem. Conventional pre-evaluated methods can get one solution only. A novel multi-objective dynamic reactive power optimization model is created in this paper and, for solving the model, a relevant improvement has been made in the non-dominated sorting genetic algorithm II (NSGA-II). This approach has been applied to the IEEE 14 nodes and 30 nodes systems and an actual system. The result shows that the approach has encouraging performance in dynamic reactive power optimization with objectives of minimizing reactive power loss, voltage deviation, and the number of operations as well.