A complex experimental uncertainty analysis method combining particle swarm optimization (PSO) and least square support vector machine (LSSVM) model was presented. Taking the tensile test of the automobile seat belt as the object, the main influencing factors and probability density function parameters of the tensile test of automobile seat belt were studied. The Latin hypercube method was used to design the test, and the finite element simulation of the tensile test of the seat belt was carried out. PSO-LSSVM was used to establish the mathematical model of the safety belt tension test, and compared with the mathematical model established by back propagation （BP） neural network. The results show that the PSO-LSSVM mathematical model hashigher prediction accuracy and mets the requirements of the subsequent uncertainty evaluation. The Monte Carlo method was further used to evaluate the uncertainty of the safety belt tensile test, and the method was compared with the method specified in the national standard. The results show that the proposed method can be applied to the uncertainty analysis of complex experiments.