Abstract: The paper designs a sum of square function, which is an optimization mathematical model that makes the residual sum of squares of wax deposition rate between the calculated and the measured minimum. The model is an unrestricted nonlinear programming problem. The author adopts Levenberg-Marquardt (L-M) optimization method which can fast and accurately seek the unknown variable of nonlinear model to compute regression parameters of the model. The effect of regression parameters using L-M optimization method seeking model is better than stepwise regression analysis method, after the algorithm exampie and the comparison.