The article presents two models for predicting cost overrun in construction projects. The first model uses regression analysis and identifies 11 key factors that lead to cost overrun, such as project size and material cost increase. The second model, based on case-based reasoning, uses past experience to estimate cost overrun. The study found that the regression model has higher prediction accuracy than the case-based reasoning model. By using the absolute value of standardized coefficient as attribute weight, the prediction accuracy of cost overrun percentage is improved.