Theoretical improvement and practical application of table-rotating method to optimize fuzzy control rules are proposed in this paper. The theorem is developed that one m×n rule table need to rotate 2(m+n−2) times at most to get all the rule tables, which are the different combinations of all the m×n×q rules. Then the fuzzy controller based on improved fuzzy table-rotating optimization is designed and applied to the systems with time delay. The simulation results show the feasibility and validity of the fuzzy controller based on the improved table-rotating rules. Compared with the conventional PID controllers and Smith predictor controllers, the advantages of this optimizing rules approach are less overshoot, shorter setting time and better robust.