Abstract: A model for predicting abundance of the Russian wheat aphid (RWA)，Diuraphis noxia (Mordvilko)，was derived from regression analysis between the actual occurrence of RWA observed and the 16 weather parameters，1989-1996 in Tacheng，with screening of the parameters step by step. The model can be expressed as: logY=8.41-0.1033RH₅ - 0.0253R₅，where Y is the number of RWA on 100 wheat tillers，RH₅ and R₅ are the relative humidity (%) and rainfall (mm) in May respectively in the year to be predicted. The model is simple and practical，and showed reliability to certain extent when used to predict occurring levels of RWA in 1997 and 1998.

Key words: Russian wheat aphid (Diuraphis noxia (Mordvilko)), occurrence level, prediction, climate factors