Abstract: Tree height prediction models were set up based on successive regression analyses on the correlation of plant height in 3 types of Pinus kesiya var.langbianensis forests and the soil factors.The height predictionmodel for the primitive forest of Pinus kesiya var.langbianensis was as:Y=0.141 519 X₁0.040 696X 5-0.047 48X₆+91.407 69;the model for the 2nd generation of Pinus kesiya var.langbianensis forest was:Y=0.039 240 9X₁+0.010 129 X₅-0.047 48 X₆+83.801 79;and the model for the artificial plantation of Pinus kesiya var.langbianensis was as:Y=0.232 25 X₁0.006 735 X₅-0.006 58 X₆+3.608 46.It was found out that the soil depth (X₆)was closely correlated to height growth of Pinus kesiya var.langbianensis trees (dependability=99%), and the altitude (X₆)and effective N content(X₁)were also closely correlated to the tree height (dependability=95%).The altitude was negatively correlated to height growth of Pinus kesiya var.langbianensis trees.Except for the tree height prediction, these models might also be used to evaluate the site degradation for Pinus kesiya var.langbianensis stands.It was showed by the study that the productivity of rotationally planted Pinus kesiya var.langbianensis forests dropped at the beginning and increased afterwards.