To improve the regulatory performance of Wellon and Edgar＇s Generalized Analytical Predictor (WE-GAP), a novel algorithm is developed to predict the load disturbance＇s effect on the process output through the disturbance response tendency. Meanwhile, a corrective parameter is introduced to improve the predictor precision and the system robustness. The corrective parameter is optimized on-line to adapt to different load disturbances. This improvement removes the assumption of WE-GAP that the disturbance transfer function should be known exactly. Additionally, the disturbance predictor can also be obtained even though the disturbance transfer function order is higher than second-order. Simulation results indicate that the proposed predictive scheme is able to improve the properties of WE-GAP and shows good control performance for chemical process with large dead time.