The transformation of scientific and technological innovations into practical applications through patent recommendation is of great significance for realizing the economic value of science and technology and promoting socio-economic development. However, existing patent recommendation algorithms often overlook the multimodal information embedded in patents, leading to recommendation results that fail to comprehensively reflect the value and application potential of patents. Consequently, this affects the accuracy of matching patents with the needs of companies.To address this issue, we propose a novel patent recommendation algorithm based on a Multimodal Heterogeneous Graph Network (MHGN).Our approach first utilizes pre-trained models to initialize the representation of the multimodal information, including the textual and image attributes of patents as well as companies information.Then, a graph attention network is employed to learn the preference representations of companies across different modalities.Based on this, we further learn the relationship weights of company-patent interactions based on the similarity of preference representations, and design a graph convolutional network to learn the node preference representations of companies and patents.Finally, to better integrate the multimodal information, we introduce an adaptation vector and use an attention mechanism to integrate the multimodal features.Additionally, we construct four real-world patent datasets from university-to-company transfers and conducted experiments comparing our model with seven advanced baseline models.The results demonstrate that our model significantly outperforms the baselines across all evaluation metrics.We release both the datasets and the source code of our model, providing robust data and model support for future research in patent recommendation and the transformation of scientific innovations.