Estimation distribution of air pollution is necessary to determine the level of pollution in a location so it can be used to recommend emission minimization. This research use some modification of Kalman Filter method for estimation distribution of air pollution. Three estimation methods namely Fractional Kalman Filter (FKF), Ensemble Kalman Filter (EnKF), Unscented Kalman Filter (UKF). The simulation result of three methods will be compared. Distribution of carbon monoxide and nitrogen dioxide in Surabaya was estimated using three methods. This reserach showed that Fractional Kalman Filter method is the best estimator among the others because error value is the smallest.