With the integration of international financial markets, the links between stock markets have become closer. Firstly, based on one-minute high-frequency returns, this paper applies the thermal optimal path (TOP) method to examine the lead–lag dependence between CSI 300 index and HSI index from 2016 to 2020. Secondly, regression analysis and correlation test are applied to cross-verify the results of TOP method. Finally, the robustness is tested through analysis with different T values and various market conditions as well as the replacing of proxy variable. The empirical results show that Hong Kong stock leads Chinese mainland stock for about one minute, and this leading effect is magnified when the stock markets fall. The experimental result is robust and has passed consistency test. This research is of great significance that not only guides investors, but also provides empirical evidence and effective information for policy makers.