The advances of single cell sequencing technologies have greatly improved our understanding of heterogeneity in terms of genetic, epigenetic and transcriptional regulation within cell populations. We and other groups have developed single-cell whole genome, exome, methylome and transcriptome technologies and applied these approaches to analyzing the complexity of cell populations in tumorigenesis, developmental process and cellular reprogramming. Meanwhile, single-cell epigenome techniques including single cell ChIP-seq, ATAC-seq, DNase-seq and Hi-C, have been developed to decipher histone modifications, chromatin accessibility landscapes, and 3D chromatin contacts respectively in single cells. Integrative analysis of single-cell multimodal data is critical for accurate dissection of cell-to-cell variation within certain cell populations. Recent progress on measuring multi-omics in the same cells has enabled analysis of associations between different layers of regulation on gene expression. So far, the relationship between chromatin accessibility and gene expression has not been investigated at the resolution of one single cell. To address this question, we report scCAT-seq, a technique for simultaneous assay of chromatin accessibility and transcriptome within the same single cell. By applying scCAT-seq to different cancer cell types, we identified trans-factors as bridges linking accessibility variation of cis-regulatory elements to cell-type-specific gene expression across single cells. We further characterized subpopulations within cancer cells and uncovered the regulatory clues that drive transcriptional heterogeneity. Together, scCAT-seq is a promising tool for the joint analysis of multimodal data of single cells, which also offers the potential for clinical applications such as preimplantation screening and cancer diagnosis.