In this research project was proposed to identify the characteristics of a competency-based curriculum, and the requirements necessary to identify the needs that must satisfy an adaptive curriculum, so it helps a student to choose the best option among a number of possibilities of offering online courses. Based on this scenario, necessary knowledge was acquired to develop a prototype, using a competency-based ontology to infer new knowledge through a semantic network and a set of predefined rules, allowing a student to generate a consistent course according to his/her skills (to be, to do, and to know-how). Somehow ensuring better performance and understanding of the issues addressed or addressed in the course. Furthermore, the result of inference indicate whether a student is relevant or not to take a course. The prototype implementation was performed using IMS-CP packages, metadata records using ontologies based on description of competencies at random, with tools such as Protege_4.0.2 version for the design and construction of ontology, NetBeans IDE 7.0 to create and programming, Apache Tomcat 7.0.11 as web server, Mysql as database manager and Mysql-connector-java-5.1.17 to connect to the database. With the development of this project is to partly solve the disadvantage of many existing platform having a commercial inclination, since the course content generated is based on a specific and rigid subject, which in most cases it’s not adapted to the requirements of each student, thus learning processes do not develop in the best way, which generate a high dropout rate.