Abstract: A method for designing the lens for uniform illumination with extended source based on natural selection particle swarm optimization (SelPSO) algorithm was proposed.The point light source lens profile was treated as the optimized object by this method. Firstly, we established a mechanism of exchanging information between Matlab and TracePro based on dynamic data exchange (DDE) technology.The characteristic advantages of both Matlab and TracePro were combined by the mechanism, so that we could construst the optimzation model easily. Then, we used the SelPSO algorithm to optimize system globally until the optimization results met the illumination requirements. In order to verify the feasibility of the system, we designed a compact uniform lens whose h/d was equal to 2.5:1. The results show that when the SelPSO and PSO are used to optimize system under the same parameters, it is found that SelPSO has faster and better global convergence capability, and its uniformity and efficiency can reach 87.566% and 89.536%,respectively. After lens installation tolerance analysis, the transverse and longitudinal displacement errors should be controlled between 0 mm and 2 mm so as to get a better lighting effect. Additionally, the method can be used for optimizing lens systems in different illuminance ranges with good illuminance effect.