Recently, nature-inspired metaheuristic optimization algorithms such as Artificial Bee Colony Algorithm (ABC) is developed. ABC is based on the feeding behavior of bee herds. ABC can not solve for time-varying function. In this study, we offer a new ABC for time-varying function. We propose ABC in which the scout bee is improved probability by normal distribution. We compare the best solution with ABC, previous method and the proposed method. Object function which optimal solution moves circle of shape is used. We investigate characteristic of proposed method according to variance of normal distribution. Best value and orbit of solutions for proposed method are better than those of other method.