We developed a database, scATAC-Ref (https://bio.liclab.net/scATAC-Ref/), to provide a large-scale resource of scATAC-seq data with known cell labels in five species. The current version of scATAC-Ref documents > 400 cell/tissue types manually annotated by published articles, covering 1,694,372 cells across five species: Human, Mouse, Danio rerio, Zea mays and Arabidopsis thaliana. Furthermore, comprehensive genome-wide chromatin accessibility analysis for known cells/tissues is crucial for exploring cell/tissue heterogeneity, and can help uncover key cell subtype and potential markers. Hence, for these scATAC-seq data with known cell labels, we used a uniform system environment and software parameters to calculate gene activity score, TF enrichment score, differential accessibility region (DAR) and pathway/GO term enrichment. The scATAC-Ref also provided a convenient, user-friendly interface to query, browse and visualize cell types of interest, to help elucidate cell type related functions and potential biological effects.