Mesoscale eddies play a significant role in marine energy transport, marine biological environment and marine climate. Due to their huge impact on the ocean, mesoscale eddy detection has become a hot research area in recent years. Therefore, more and more people are entering the field of mesoscale eddy detection. However, the existing detection methods mainly based on traditional detection methods typically only use Sea Surface Height (SSH) as a variable to detect, resulting in inaccurate performance. In this paper, we propose a mesoscale eddy detection method based on multivariate fusion data to solve this problem. We not only use the SSH variable, but also add the two variables: Sea Surface Temperature (SST) and velocity of flow, achieving a multivariate information fusion input. We design a novel symmetric network, which merges low-level feature maps from the downsampling pathway and high-level feature maps from the upsampling pathway by lateral connection. In addition, we apply dilated convolutions to network structure to increase the receptive field and obtain more contextual information in the case of constant parameter. In the end, we demonstrate the effectiveness of our method on dataset provided by us, achieving the test set performance of 97.06% , greatly improved the performance of previous methods of mesoscale eddy detection.