As there exists almost no convergence or performance evaluation theory for nondeterministic polynomial (NP) time algorithms, illustrating the graphical method has become a popular tool in comparing the convergence of computational intelligence algorithms, but the comparison is inconclusive if the curves cross multiple times. To improve, this paper proposes a method to evaluate whether optimization algorithm obtains the best possible solution with the least possible computational costs. This is achieved by adding a computational cost weight on the sequence of the historical best solution. To validate this method, we carry out experiments on evolutionary algorithms with respect to the optimality performance. The results show that this time-weighting method evaluates the convergence performance more effectively and directly, revealing not only the convergence speed but also whether the algorithm finds the global optimum on benchmark functions.