Abstract: Based on research of total power of agricultural machinery in Lvliang city from 1979 to 2015，the exponential function，the cubic polynomial function and the BP neural network were used to establish prediction model of total power of agricultural machinery and compare the samples.Results showed that average absolute errors of the BP neural network and the exponential function model were 1.11% and 3.22%，respectively，which were lower than average absolute error of the cubic polynomial function （8.05%）.Finally，the BP neural network model and the exponential function model were used to predict total power of agricultural machinery in Lvliang city in 2016—2021，with a view to providing a reference for development of agricultural mechanization level.