In this paper, a constrained mean-variance model is constructed for the portfolio optimization problems. The model is a mixed quadratic integer programming problem, and it is too hard to solve by using the traditional optimal algorithms. The purpose of this paper is to use a heuristic algorithm to solve this problem. Combined with the differential evolution strategy, a new hybrid artificial bee colony algorithm is proposed for solving the constrained mean-variance model. In the new hybrid artificial bee colony algorithm, a new search scheme and obsolete rules are presented to improve convergent speed of the algorithm. Computational results show that the algorithm has great capability to solve the portfolio optimization problem. It supplies a new method for solving the portfolio optimization problem.