This paper explores the potential of machine learning algorithms (MLAs) for the simulation of intercity networks. To this end, we implement the random forest MLA to simulate the intercity corporate networks created by Fortune China 500 firms in mainland China. The random forest MLA does not require a predefined model but detects patterns directly from the data to automatically build models. The city-dyad connectivities were computed using an interlocking network model and treated as target variables. City factors and geographical factors were treated as features. The model was trained using a 2010 training set and subsequently validated using 2010 and 2017 test sets. The results are promising, with the pseudo R² of the model coupled with different test data ranging from 0.861 to 0.940. Nonetheless, the random forest MLA also faces some challenges in the context of the simulation of intercity networks. We conclude that MLAs are potentially useful for specific applications such as the analysis of network big data, scenario simulation in regional planning, and the detection of driving forces in exploratory research.