Abstract: The proliferation of online rumors and their negative social impacts demands efficient rumor detection mechanisms. Considering the fact that astroturfers often appear in the process of rumor propagation, it is feasible to use the number of astroturfers to assist in rumor recognition based on the characteristics of forwarding users. At the same time, the emergence of debunkers is also important for determining rumors, mining debunker features as one of the forwarding user features. This paper proposes a Microblog Rumor Detection Model Integrating Forwarding User Features (MRFUF) to improve the accuracy of rumor detection. This paper employs the BERT pre-trained model to extract textual features and integrates user characteristics using the Multimodal Adaptation Gate (MAG). The effectiveness of the MRFUF model is validated through joint experiments on a microblog dataset and an astroturfer dataset.