The explosive growth of information on the internet poses great challenges to users. In response to the above issues, this article studies the construction of an intelligent recommendation system for internet information using a combination of deep learning and big data methods. This article adopts deep learning methods to model user behavior data and combines big data analysis to provide personalized recommendation services for users. This article compared the information recommendation accuracy of two other intelligent recommendation systems through experiments. The system designed in this article achieved an accuracy of 92.5% in this regard, far superior to the other two. Experimental results have shown that the system designed in this article not only improves recommendation accuracy, but also enhances user satisfaction.