Abstract: In this paper a method is proposed, which is based on my previous work"Gradient Acceleration Method for Solving Unconstrained Optimization Problem".It hasnot only some advantages in original scheme. but also appears more effective in accele-rating its convergence near the optimal point.