Abstract: In view of the periodic characteristics of present emission reduction targets, the control of pollutant emission is only needed to meet certain requirements of emission reduction while not necessary to reach the minimum value. By taking pollutant emission as environmental constraints, the dynamic stable economic scheduling model is built by considering the unit efficiency problem at the same time. Through choosing different fluctuation coefficients and fluctuation functions, the purpose that units with different capacity can bear different fluctuations is achieved. The improved multi-objective particle swarm optimal algorithm is applied to solve the model by introducing the maxmin function and ε dominate factor. The example shows that an operation scheduling scheme can be obtained from this model by taking economy and stability into account under given environment constraints, which can guarantee output stability of large-capacity units and improve output curve of small-capacity unit.