In order to meet the traffic demand and improve the effective of the traffic signal control under uncertain condition, this paper initially proposed a new signal timing interval optimization model of traffic signal control based on the interval optimization theory. Firstly, the traffic flow interval in rush hour was determined by 5 min acquisition section and used as the input parameters of traffic control system. Secondly, the interval-valued nonlinear optimization model of signal timing was constructed by taking the average vehicle control delay interval as the objective function.Thirdly, taking the intersection service level as the performance objective, this model was transformed by using the interval possibility degree model and solved by using genetic algorithm. Finally, the three-phase and two-phase intersections in rush hour in Beijing were simulated by using VISSIM simulation software. The results show that the method is feasible and effective, and is obviously superior to Webster scheme, robust optimization method and measured data at three-phase intersection with large fluctuation of traffic flow. The vehicle average delay decreases by 26.57% and capacity increases 7.07% compared with the measured data. At the two-phase intersections with relatively stable traffic volume, the interval optimization method is feasible and can achieve the expected control effect, but it is weaker than other schemes. Furthermore, the signal timing interval optimization model was used by the eight intersections in rush hour in Beijing with VISSIM simulation software and analyzed by the fluctuation trend of the saturation and the average delay of the 5 min acquisition section.The analysis show that the interval optimization model can achieve the expected control effect for each intersection during peak hours, and is more suitable for intersections with large traffic flow fluctuations, and the performance of the control scheme is more stable.