Dynamic Alliance Partner Selecting problem is generally modeled as a combination optimization problem, the popular solution of this problem is building math models and choosing proper optimization method. Researchers have already brought forward many useful models and algorithms, but till now no solution of Partner Selection Problem is public accepted due to the problem's complexion, which means there is a need to search for better Partner Selection methods. ACO (Ant Colony Optimization Algorithm) is a useful method in solving combination optimization problems, however, it is seldom used to solve Dynamic Alliance Partner Selection problem, and so relative researches are needed to test whether ACO can be a new method in solving Partner Selection Problem. In this essay, an ACO-based Dynamic Alliance Partner Selection Model is brought forward, and simulation is done to prove the validity of the model, this means a lot to the research of both Dynamic Alliance Partner Selection Problem and ACO.