Abstract: 【Objective】The individual tree diameter growth model of a Larix olgensis plantation in Heilongjiang Province was developed by a single level linear mixed model. The model had good prediction accuracy, and it provided a theoretical basis for accurately predicting the growth and reasonable management for Larix olgensis plantations in Heilongjiang Province.【Methods】Based on the data of 148 fixed plots in Heilongjiang Province, this study used a stepwise regression method to introduce the initial size factor, competition, and site factors for trees, and established and evaluated five different independent variables(5-year later diameter(d₅), a periodic annual increment model using 5-year diameter increment(d₅- d₀), the natural logarithm of 5-year diameter increment [ln(d₅ - d₀ + 1)], the natural logarithm of 5-year squared diameter increment [ln(d²₅ - d²₀ + 1)], and 5-year squared diameter increment(d²₅ - d²₀)). Additionally, a set of generalized distance-independent individual tree diameter growth models were constructed by using the dummy variable method based on the optimal traditional model. In addition, the individual tree diameter growth model was built based on a single level linear mixed model with a parameter random effect for plot on the basis of the dummy variable model. Each approach was evaluated by using independent validation data.【Results】We could clearly find the progressive improvement in AIC when competition and site were added in a stepwise regression as predictor variables for individual tree diameter growth model with each response variable. The model of ln(d₅ - d₀ + 1)for response variable was the best model for the individual-tree diameter increment prediction. The natural logarithm of the initial diameter(ln d₀), total stand basal area per hectare, basal area of the trees higher than the object tree, and site index were found to be significant predictors. The dummy variable models showed the regional difference while ensuring prediction accuracy. The R², MSE and RMSE of the mixed effects model were 0.978 3, 0.713 7 and 0.844 8 cm, respectively. Compared with the traditional model, the MSE and RMSE of the mixed model were reduced by 0.300 6 and 0.162 3 cm, respectively, and the R² values were practically equivalent. The linear mixed model, including fixed and random parameters, provided a better fit among the models tested.【Conclusion】Compared with the traditional model, the individual-tree diameter growth model for the Larix olgensis plantation in Heilongjiang Province based on the linear mixed effect improved the prediction accuracy, which provided a reliable theoretical basis for accurately predicting the growth of Larix olgensis plantations in different regions of Heilongjiang Province.