Abstract: The occurrence and development of cancer are highly heterogeneous. Effective integration of multi-omics information to accurately identify cancer subtypes is the key to achieve precision medical treatment. Therefore, a method based on multi-kernel network fusion of multi-omics data to identify cancer subtypes is proposed, which can be used to effectively process high-dimensional and non-linear data. For each omics dataset, a different kernel function was used to calculate the similarity between the samples, and the resulting similarity matrix captured the underlying relationships between the samples in each omics layer, enhancing a comprehensive understanding of the cancer subtype. Comprehensive experiments demonstrate that the proposed method obtains more significant results than the eleven methods on six datasets in ten cancer datasets.