In order to further study the growth model between DBHand tree height of Picea schernkiana in xinjiang, fill the gap in related research , and provide a useful reference for the growth prediction and production and management of P. schernkiana stand. In this paper , 30quadrats , each with a size of 25m ×25m , were used to construct the tree height-DBH growth model of Tianshan spruce by selecting 34 kinds of growth equations , and it was found that the Logistic H2 model H=1.3+ a /(l+b-lD-c)and the Mixed model H=1.3+ a /(l+b-lD-c)had the best fitting effect. The adjusted R2 was 0.862m , the RMSE was 4.768m,the MAE was3.625m,and the SSE was l0, 839.76m. In addition, a nonlinear mixed-effect model was constructed by adding the most random effects of altitude and slope in this study , and compared with AIC, BIC and log-likelihood values,it was found that the nonlinear mixed-effect model based on altitude had better than effect and higher prediction accuracy.