Tropical cyclones (TCs) cause significant damage to coastal communities and infrastructure each year. Probabilistic coastal hazard assessments provide information on the frequency and severity of coastal hazards (typically represented as hazard curves) and serve as a key input to probabilistic risk assessments for coastal hazards. The Joint Probability Method (JPM) serves as a foundational tool for the probabilistic assessment of TC-induced hazards, especially for storm surge. Among the critical TC parameters used in the JPM, including central pressure deficit, forward velocity, radius of maximum wind speed, and heading angle, the heading angle is a circular variable and represents the direction of TC approach. It exhibits statistical properties that differ fundamentally from those of linear parameters. Existing studies have begun to explore the impact of incorporating circular statistics into the modeling of the heading angle within the JPM framework for regional case studies. However, the circular statistical characteristics of the heading angle across multiple geographic regions have not been systematically investigated. In this study, we investigate the circular statistical properties of the heading angle across multiple study regions using different TC data sampling methods. We focus on the variability of the heading angle’s marginal distributions and its statistical dependence on other linear TC parameters. The results provide new insights into regional variability and dependence structures among TC parameters and highlight the importance of properly accounting for circular variables in probabilistic TC hazard modeling. These findings contribute to improving the accuracy and robustness of JPM-based assessments of TC-induced coastal hazards.