Abstract:At present,the algorithm of multi-objective SVM model which uses multi-threshold is not ideal in the efficiency and computation.This article proposes the SVM model using the same threshold value,and introduces the depth-first search algorithm to the process of SVM parameter optimization,and then uses the sequential minimal algorithm to improve the SVM features.Results show that parameter optimization using depth-first search algorithm is relatively good.The model is an effective method for multi-objective problems.