In recent years, with the rapid development of online vocational education, how to achieve personalized learning recommendations has become a key issue that needs to be addressed. This paper proposes a collaborative filtering model integrated with the Nonlinear Conjugate Gradient (NCG) algorithm and applies it to the personalized design of vocational education online learning platforms. By improving the similarity calculation and optimization strategies in traditional collaborative filtering algorithms, the recommendation accuracy and efficiency are enhanced. This paper first analyzes the user behavior and data features of vocational education online learning platforms, constructs an NCG-based collaborative filtering recommendation framework, and carries out its system design and implementation. In the experimental section, results based on real datasets show that the proposed method can effectively improve the recommendation system"s accuracy and efficiency, providing efficient personalized service support for online vocational education learning.