To present the first syntaxonomical classification for the mesophilous deciduous forests of the Pamir-Alai Mountains in Tajikistan with some remarks on its environmental gradients. Location: Tajikistan. Methods: Altogether 201 relevés were sampled between 2008-2013 using the seven-degree cover-abundance scale of Braun-Blanquet. They were classified by the modified TWINSPAN method using the four-step interval scale with cutoff levels of 0%, 2%, 5% and 10% and total inertia as a measure of cluster heterogeneity. Diagnostic species were identified using the phi coefficient as a fidelity measure. Detrendended Correspondence Analysis was used to determine the relation between samples, vegetation units and the major gradients in species composition. Results: Plant communities of typical broad-leaved, riparian and gallery woods inhabiting mainly the lowland, colline, montane and subalpine zone in several mountains and river valleys in the Pamir Alai Mts. are herein described. Numerical analyses of vegetation data resulted in the distinction of ten associations: Juglandetum regiae, Aceretum turkestanici, Violo suavis-Populetum albae, Swido darvasicae-Platanetum orientalis, Populetum pamiricae, Populetum talassicae, Salicetum turanico-pycnostachyae, Populetum pruinosae, Betuletum turkestanico-tianschanicae and Fraxinetum sogdianae, which have been assigned to four alliances: Acero turkestanici-Juglandion regiae for mesophilous broad-leaved forests, Populion afghanicae for montane riparian forests, Populion talassicae for subalpine riparian forests and Elaeagno-Populion pruinosae for gallery forests of the Middle Asian lowlands. The main factors determining the species composition of the studied associations are salinity, soil humidity, elevation, slope and aspect. Conclusions: Our first attempt initiates research on the diversity of forest vegetation in Tajikistan and Middle Asia. It provides the first classification scheme that could be used in further studies on forest vegetation in the surrounding regions, particularily in Middle and Central Asia.