In order to effectively improve the accuracy of dynamic control of construction cost, the ant colony optimization algorithm and fuzzy Petri net theory are adopted to predict the construction cost. First, the sample project is selected by the rules of fuzzy production and the similarity between the projects is determined so as to establish the project cost forecasting model. The weights and thresholds are trained by BP neural network. Then, the parameters of the model are optimized by using ant colony optimization, so as to further improve the accuracy of project cost forecasting. The results of actual construction project show that compared with the traditional BP neural network prediction method, the proposed method has higher accuracy and can be effectively applied to the scientific management of construction cost.