During the internet era, multimodal content plays a significant role, and the interaction of users with multimodal items remains a key focus in recommendation system research. Multimodal recommender systems that utilize multimodal features, such as textual descriptions and images, tend to offer more accurate recommendations compared to traditional recommender systems that depend solely on interactions between users and items. Current multimodal recommendation models focus primarily on item-item or user-user relationships, neglecting the potential benefits of combining both to enhance the connections between users and items. Moreover, these models often overlook the exploration of higher-order relationships among items or users. Building on the existing multimodal recommendation model FREEDOM, we introduce a new multimodal recommendation model enhanced by a user isomorphic graph, named UCEMR. Specifically, UCEMR efficiently considers user relationships, introduces a novel user similarity calculation formula to uncover deeper connections between users, and employs a straightforward fusion method to derive user and item representations effectively. Evaluation of the proposed UCEMR model across three practical datasets reveals its superior performance compared to current leading baseline models.