The Internet database for food composition in Taiwan, including food descriptions, sample descriptions, analysis items, methods, and references, was built by the Department of Health. Although complete food composition data has been collected in this database, it is not convenient for people who do not have sufficient nutrition knowledge to retrieve nutrition information. Recently, diet management is an important topic in health care. To provide sufficient nutrition information, the recipe ontology was built based on common food recipes in this study. This ontology was used to explicitly define terms of food composition and design analytic procedures and search functions. In this recipe ontology, recipes were classified as 7 categories and foods were classified as 18 categories referred to the database for food composition in Taiwan area. The recipe ontology was implemented using XML and integrated with MySQL and ASP to develop a food composition analysis system. In this system, users can retrieve nutrition information and evaluate their dietary intakes according to the 24-hour dietary record. In addition, the system can generate dietary recommendations according to the personal records. The food composition analysis system was evaluated by the experts in nutrition using a questionnaire survey. According to the questionnaire results, more than 80.96% experts consider that the system is helpful for education, research, and clinical application. Therefore, this system can help people to complete the goal of diet self-management for health.