In order to improve positioning accuracy of GPS/DR integrated navigation, the fuzzy adaptive Kalman filter method was analyzed. Experiment was carried out on the modified agricultural machinery. Experiment results showed that the fuzzy adaptive Kalman filtering could inhibit the influence of unusual values on the localization precision of the system, and the system positioning accuracy of the proposed method was better than single GPS. The average positioning error of the x and y directions were 0.13m and 0.20m by using the fuzzy adaptive Kalman filter. Positioning curves was smoother after filer.