Hyrcanian forests in northern Iran are one of the world's precious deciduous broadleaved forest ecosystemsbut are subjected to increasing disturbances mostly of anthropogenic origin.In this context, it is necessary to assessthe existing vegetation and its relation toenvironmental factors in order to establish a scientific reference for conservation measures.The purpose of this research is to evaluate the quality of the modified and classical TWINSPAN (Two-way indicator species analysis) classification based on the environmental factors toidentify the plant communities of Taxus baccataL. forests of Golestan Province (Northern Iran). Fifty400-m2 plots were systematically selected along an elevation gradient in the Taxus baccataforests. In each plot, the soil was sampled usinga compositesample of threereplicates taken to a depth of 20 cm.All vascular species were recordedon each plot according to Braun-Blanquet’s method. In total,56 plant species belongingto 48 genera and 34 families were recordedin spring 2018. The comparison of the results of modified and classicalTWINSPAN showed that the modified TWINSPAN did not change the classification logic, but introduced moreflexibility in the hierarchy of the division.It was also found that the compliance rate of the results provided by the classical TWINSPAN method was lower than the modified method. Usingmodified TWINSPAN, Detrended correspondence analysis (DCA), and Redundancy analysis (RDA),we determined four floristic groups in the Taxus baccatahabitat.The first(Carpinus betulus) and second (Taxus baccata) groups were located at low altitudes and soil contents in silt and clay were higherthan in the other groups. In contrast, the third (Quercus castaneifolia) and the fourth (Juniperus communis) groups were located next to each otherat higher altitudes and southwestern aspects, on steeper slopesand soils that exhibited a highpH and a high sand content.We foundclose relationshipsbetween the distribution of vegetation groups of Taxus baccatahabitat withedaphic and physiographic factorsusing the RDA method.To conclude, the modified TWINSPAN offers a more efficient and flexible method to identify vegetation communities. This is thefirst step to implementingfuture management and ecological restoration measuresin habitats of high ecological value such as yew forests.