A robust algorithm of weighted least squares support vector machine(LSSVM-W) was developed to reduce the effect of outliers on DEM construction. LSSVM-W used the result(parameters array α) of the LSSVM as the initial values to compute the weights of sampling points. Then, a iteratively LSSVM-W was formed to compute a more perfect parameters array. A synthetic surface was employed to comparatively analyze the estimation accuracies of LSSVM and LSSVM-W,where the sampling points were contaminated by the standa...