As a vital source of the world’s economic development, an effective means to obtain foreign direct investments (FDI) in emerging markets is of vital interest in economic development. But it is hard for traditional evaluation methods to express the effects of FDI adequately. While intelligent computing systems such as AI and machine learning can deal with more complex massive data, establish specific evaluation models, and improve the accuracy and effectiveness of evaluation. This paper constructs the theoretical framework for intelligent computing evaluation of effects of foreign direct investment of emerging markets, and puts forward three research hypotheses that intelligent computing technology can significantly improve the accuracy and make the evaluation convenient and effective, and market characteristics and regulatory conditions constrain its application. This paper performs empirical research based on constructing an intelligent computing evaluation model through the random forest (RF) algorithm and comparing it with the traditional algorithm. The conclusion shows that the intelligent computing evaluation model is better than the traditional method in error statistics, and verifies the research hypotheses. The study provides the scientific foundation for FDI policies in developing market countries and guides the future direction of optimization of models’ applications, such as feature selection optimization and model interpretation optimization.