In recent years, shared cycle services have expanded in many countries, including Japan. Understanding the needs of users is essential for the future development of shared-cycle services and for improving their usage. On the other hand, it is difficult to collect user needs in Japan, where there are many unmanned shared-cycle services. In this study, we constructed a model to estimate the purpose of use by combining actual use data, questionnaire data, and building area data around ports, and applying machine learning models, Random Forest Model and XGBoost Model, to the model. The results of interpreting each model using the SHAP indicator indicated that travel distance and travel start time were important explanatory variables in determining the purpose of use. In addition, we applied the models to the actual usage data for multiple years to understand changes in user behavior.