The modified two pilot behavior models are discussed. One of them is the structural model. Its modification allowed to get better agreement in calculation of variance of error and to evaluate the influence of some new task variables (control element gain coefficient, requirements to the accuracy) on pilot – vehicle system parameters. The other model is the composite model of pilot behavior based on neural network approach. This model provides high agreement between the

simulated and experimental pilot frequency response characteristics. The structural model is used for development of the criterion for the of flying qualities prediction in lateral channel with taking into account the influence of motion cues on pilot behavior and for the preliminary design of the predictive display. As for the composite model it was used for development of the flying qualities criterion for the pitch control task based on calculation of pilot and pilot aircraft system frequency response parameters. The additional procedure — the preliminary selection of dynamic configurations from the database was proposed what allowed to get high agreement between the predicted flying qualities level evaluated by the pilots in flight tests.