Abstract: To improve the efficiency of signal timing plans for urban traffic roads,an automatic generation algorithm which combines micro traffic environment modeling and signal optimization is proposed here.In this method,the image processing technique is utilized to implement automatic extraction and construction of micro-traffic simulation environment by combining the characteristics of urban traffic and raster map.Then the improved car-following and lane-changing model is employed to model the behaviors of urban traffic vehicles dynamically.Based on that,the improved fast non-dominated sorting genetic algorithm(NSGAII)optimization model is put forward,which takes green time ratio,phase sequence,offset and period as optimization parameters,and the average delay,average queuing-lane length ratio and the system capacity as optimization targets.Ultimately,the experimental results from arterial network with multiple intersections have been given to validate the solving efficiency and control performance of the proposed algorithm.