To increase the precision of high pressure common rail forecast model, the modeling of high pressure common rail diesel engine based on AMESim was introduced. On this basis, grey relational theoretical analysis was used to analyze the multi-parameter system and calculation to determine the input and output variables of the predictive model. Adaptive weighted Particle Swarm Optimization algorithm was applied to the optimization of initial parameters of least square support vector machine. Through the examination of 20 forecasting samples, the maximal error of the forecast model is 0.079 1, and the average relative error is reduced to 0.039 6 by the least square support vector machine, which is far superior to commonly used empirical formula and neural network.