It is generally expected that inter-annual changes in radial growth among trees would be similar to the increase in altitude due to the limitation of increasingly harsher climatic factors. Here, we examine whether this pattern exists in alpine forests on the southeastern Tibetan Plateau. Increment cores were collected from mature trees at the lower, middle and upper limits of balfour spruce (Picea likiangensis var. balfouriana (Rehd. et Wils.) Hillier ex Slsvin) forests at the Buze and Yela Mountains in Basu County, Changdu Prefecture of Tibet, China. The treeline elevations are 4320 m and 4510 m a.s.l. for Buze and Yela, respectively. Tree-ring widths were measured, crossdated, and detrended to obtain a sequence of ring-width indices for each individual sample. Annual growth rate, climate sensitivity, growth-climate relationships, and growth synchrony among trees were calculated and compared across altitudes. In Buze Mountain, the annual growth rate of trees has no significant difference across altitudes. The mean sensitivity of trees is lower at the treelines than at lower elevations. Tree growth has stronger correlation with winter temperature at upper elevations than at lower elevations, has significant correlation with moisture, not temperature, in the growing season, and the growth response to moisture is lower at the treeline than at lower elevations. The correlation among individual tree-ring sequences is lower at the treeline than at sites at lower elevation. In Yela Mountain, the characterisitics of annual growth rate, mean sensitivity, tree growth-climate relationships, and inter-serial correlation are similar to those in Buze, but their differences along altitudinal gradients are less significant as those in Buze. Our data do not support the general expectation of growth convergence among individuals with increasing altitude. We conclude that individual heterogeneity and microhabitat diversity are important features for treeline trees that may dampen the growth synchrony in trees. The results obtained in this study expand our knowledge about the pattern of forest growth along altitudinal gradients in high-elevation regions and demonstrate the importance of checking the growth of tree individuals before analyzing the average signal.