With the rapid development of Internet technology, the digital contract is a vital medium of information dissemination. The rapid spread of rumors in Internet news communication misleads public cognition and may cause social panic and instability. This study aims to explore a deep reinforcement learning (DRL) model for rumor detection in news communication that integrates an attention mechanism. By enriching and advancing the theoretical framework and technical methods in the field of rumor detection, it seeks to provide more accurate and reliable information support for Internet news communication. This study first analyzes the public opinion transmission cycle of Internet news. It makes it clear that the word vector model is the key link in the natural language processing task of news rumor detection. In the task of rumor detection, a Deep Q-Network (DQN) detection model is further proposed, which takes text information as input, extracts features, and learns optimal strategies through the deep neural network, to achieve accurate recognition of rumors. The attention mechanism is integrated into the model to help it better identify the key information in the text, such as