Chatbots can change health behaviors by generating conversations about health. A well-designed health-related chatbot can encourage people to be physically active by creating a therapeutic relationship. This study aimed to develop and evaluate the feasibility and usability of a chatbot-based physical activity self-management application to promote physical activity among community-dwelling older adults. The physical activity self-management application was designed based on the five stages of the ADDIE model (analysis, design, development, implementation, and evaluation). The mobile application (app) includes a chatbot and physical activity videos as key components to provide support and guidance for physical activity enhancement and self-management. Quantitative pilot testing demonstrated that participants were highly satisfied with the app's usability, convenience, and overall satisfaction across all measures. Qualitative data analysis revealed that participants found it easy to watch and follow the videos as needed throughout the day and reported that the guidance and support from the chatbot helped motivate them to be more physically active. The findings suggest that it may be beneficial to expand the use of the chatbot-based physical activity self-management application to promote physical activity among older adults with chronic diseases and offer promising evidence for the feasibility of increasing mHealth accessibility for lifestyle modification interventions.