In recent years, with the rapid development of modern technology, people are constantly exposed to recommended new media content through media such as mobile phones. Traditional new media content recommendation algorithms often face insufficient personalization, and the recommended content is also recommended to unwanted audiences, resulting in a lack of targeted recommendations. How to improve the recommendation quality of new media content has become an urgent problem that needs to be solved. Artificial intelligence (AI) algorithm technology can achieve precise recommendations, more accurately capture user new areas, and achieve personalized recommendations, so as to ensure that recommended content is more in line with user preferences, and improve recommendation effectiveness. This article took the FNN (Feed-Forward Network) algorithm model as the research object, integrated it with the KNN (K-Nearest Neighbors) algorithm, and proposed a new ConvFNN-KNN model. It was found that the fused algorithm outperformed other algorithms in terms of accuracy, diversity, and coverage.