Abstract: A modified Support Vector Machines (Modified SVM) for non-parametric estimation of density function,which is a modified version of Support Vector Machines (SVM) for non-parametric estimation of density function,is discussed.This procedure is based on the prior knowledge that the probability of any event decides the number of samples which indicate the happening of that event in probability theory,specifically,the larger probability of an event,the more number of samples which imply the happening of that event.A better outcome is obtained by Modified SVM than that by SVM in the simulation.What’s more,since the problem of the estimation of density function is ill-posed,the regularization method is adopted.