Background and objectives: One of the main fields of interest in vegetation ecology is the analysis and understanding of vegetation-site relationships, particularly the response of species to underlying ecological gradients. Many attempts have been done in linking the performance of plant species to environmental variables since last decades. The aim of this study is investigating on the response curve of the beech tree to environmental variables using the generalized additive model.Materials and methods: For this purpose, a stratified sampling method based on landform was used to locate 114 0.1 ha circular sample plots in beech dominated forests in experimental and educational forest of Kheyrud, Nowshar. The mean height of five largest diameter trees within each plot was considered as dominant height. Elevation above sea level, geographical aspect and slope of the ground were also recorded or measured. At the center of plot, soil samples from 0-10 cm depth were taken for analyzing soil texture, bulk density and saturation moisture, pH, lime (%), nitrogen (%), carbon and organic matter (%), potassium, calcium, magnesium and phosphorous. By using generalized additive model and mgcv package in R statistical software, the response curves of the beech tree were individually and simultaneously analyzed. Due to the nature of response variable, Gaussian distribution and identity link function were selected for generalized additive model.Results: The comparison of response curves resulted from GAM for explanatory variables individually and simultaneously showed that there are considerable differences in the shape of the response curve. There are also differences in significance of predictors among these two approaches. By using GAM for each explanatory variable individually indicated that altitude, slope, radiation index, clay, silts, sand, nitrogen, saturation moisture, carbon and organic matter, pH, phosphorous and potassium are significant (P<0.05). Considering all non-collinear predictors showed that altitude, ration index, sand, bulk density, nitrogen, CN, pH and phosphorous are significant variables on beech dominant height in the generalized additive model.Conclusion: The results of this research imply that if the study aims at investigating only the shape of the response curve, considering simultaneously all predictors will present the precise description of species behavior to environmental variables. But if the researcher wants to extract ecological optimum and tolerance for the species besides the shape of the response curve, the behavior of species to predictors individually would be the better choice.