The growing number of multi-omics studies demands clear conceptual workflows coupled with easy-to-use software tools to facilitate data analysis and interpretation. This protocol covers three key components involved in multi-omics analysis, including single-omics data analysis, knowledge-driven integration using biological networks and data-driven integration through joint dimensionality reduction. Using the dataset from a recent multi-omics study of human pancreatic islet tissue and plasma samples, the first section introduces how to perform transcriptomics/proteomics data analysis using ExpressAnalyst and lipidomics data analysis using MetaboAnalyst. On the basis of significant features detected in these workflows, the second section demonstrates how to perform knowledge-driven integration using OmicsNet. The last section illustrates how to perform data-driven integration from the normalized omics data and metadata using OmicsAnalyst. The complete protocol can be executed in ~2 h. Compared with other available options for multi-omics integration, the Analyst software suite described in this protocol enables researchers to perform a wide range of omics data analysis tasks via a user-friendly web interface.