To introduce the green traffic idea and the bus priority idea into the modeling strategy of signal timing optimization for intersections, a multi-objective signal timing optimization model varying with the phase effective green light time was proposed by considering the per capita delay and per capita CO emissions as the indexes. The bus deduction coefficient was introduced into the delay per capita to overcome the negative effect of absolute priority on private car. The fuzzy compromise method was used to transform the two objective functions of different dimensions into a single objective function, and to determine the values of two dimensions lie in (0, l). Fuzzy preference method was used to determine the membership function weights in the single objective function. The improved PSO (particle swarm optimization) which combines the SAPSO (self-adaptive particle swarm optimization) and the AsyLnCPSO (asynchronous learning-factor changing particle swarm optimization) were used to solve the single objective function based on the MATLAB software platform. Finally, the model was applied to an actual case and the target values were compared and analyzed. Results showed that the per capita delay reduced by 0.94s and decreased by 3.87% after optimization. The per capita CO emission reduced by 1.25g and decreased by 12.74%. The optimized signal timing scheme has an optimal effect on delay and emission, and the validity of the model was observed.