Abstract: The ordered plot clustering, belonging to the non-hierarchical agglomerative algorithms, was introduced in the vertical belt partitioning of Mian Mountains, Shanxi in this paper. The result of the vertical belt partitioning is as follows: 1. the deciduous broad leaved forest belt (800～2000m above sea level),including 3 subbelts: [Ⅰ. the deciduous broad leaved shrub subbelt (800～1200m above sea level); Ⅱ. the pine and oak forest subbelt (1200～1700m above sea level) ;and Ⅲ. the deciduous broad leaved forest subbelt (1700～2000m above sea level)];2, the cold-temperate coniferous forest belt (2000～2200m above sea level) ;and 3, the subalpine meadow belt (2200～2440m above sea level).Based on the algorithm of the gold partitioning method(Fisher’s method),ordered plot clustering was carried out according to , firstly, the criterion of plot order from the first to last one, and secondly ,the criterion of the maximal similarity within each plot group (or the minimal similarity among plot groups). Thus, the classification is optimal. In order to compare with ordered plot clustering, TWINSPAN was applied to the vertical belt partitioning of the mountains. The results showed that the classification of the ordered plot clustering was superior to that of TWlNSPAN in the case of the serious disturbances of primary vegetation.