A method for improving local optimal solutions of nonlinear programming problems by treating constraints directly, named "Modal Trimming Method, " has been proposed by the authors. This method combines a gradient method for searching local optimal solutions, with an extended Newton-Raphson method based on the Moore-Penrose generalized inverse of a Jacobian matrix for searching initial feasible solutions used for the gradient method. In this paper, the strategy for preventing traps into fathomed local optimal solutions is revised to treat a wide range of problems. The method is applied to three types of test problems-quadratic programming, quadratically constrained, and nonlinear programming ones, and its performance is evaluated in terms of the global optimality of the suboptimal solutions obtained by the method. It turns out that the method has a high possibility of deriving the global optimal solutions by the suboptimal ones for a wide range of problems.