Labor education is an important content of quality education, and the promotion of higher education cannot be separated from the implementation and practice of labor education. The labor education platform and labor education resource base in colleges and universities are established using the big data platform, which combines Hadoop technology and the HBase database. In order to help students obtain labor education teaching resources more efficiently on the labor education platform, reinforcement learning is combined with a deep neural network to optimize the scheduling of teaching resources, and α -the dispersion recommendation algorithm is used to realize personalized recommendations of teaching resources. With regard to the effectiveness of the labor education platform, load testing and application effects were carried out, and the influencing factors of labor behavior and craftsmanship cultivation of college labor education students were analyzed in depth. The labor education platform’s transaction response time increases by 22.5 seconds when the number of users reaches 5000, according to the results. The personal motivation for craftsmanship in the process of labor education will be affected by the campus culture, and its correlation coefficient is 0.565, and the student’s satisfaction with the platform of labor education reaches 0.825. The innovation and development of labor education in colleges and universities need to fully rely on the big data platform to promote the optimization and sharing of labor education resources and to promote the development of labor education resources. The cultivation of student labor creation and craftsmanship can be ensured through the optimization and sharing of educational resources.