The the article the issues associated with the use of fuzzy logic in a non-linear Kalman filter working in adaptive control structure of the dual mass system are presented. In the introduction, a brief review of the literature is done. Then the mathematical model of the drive system is shown. Next, the control structure and the fuzzy extended Kalman filter is presented. The adaptation of the coefficients of the covariance matrix of the Kalman filter with the help of the fuzzy system is described. Estimator is tested in an open and closed control structure. Compared to the classic solutions, a significant improvement in the quality of the state variable estimation of the drive system is obtained.