To predict gear wear rapidly and accurately, a new wear numerical simulation model is established considering distributed load between double tooth meshing area based on Kriging method. The distributed pressure and meshing speed, which should be obtained before calculating wear depth, are obtained based on the Winkler surface model and gear meshing theory. The Load to determine distributed pressure is distributed dynamically considering the effect of clearance caused by wear. Based on Archard's wear model, the calculation wear model of spur gear is derived, and the wear depth of each meshing points on tooth profile under the different wear cycles are obtained. A surrogate model which can describe the relation between wear and gear parameters is constructed based on the Kriging and artificial neural network method, the approximation level and goodness of fit among different Kriging models with different samples are studied. As shown in a numerical example, the wear depth tends to accumulate as wear cycles and are varying over the teeth flanks with minimum wear at the pitch and increase as the meshing point moves towards the root. By comparison of three Kriging models with different original sample numbers, the minimum sample number is 100 if the approximation level and goodness of fit meet the demands. The Kriging model has high computational efficiency and accuracy and can overcome time-consuming defect.