Single-cell omics technologies are promising tools for disease diagnosis, treatment and prognosis. While single-cell genomics and transcriptomics are blossoming, in-depth and large-scale research into single cell proteomics is rare due to technical limitations. In this project, we established mass-adaptive coating-assisted single-cell proteomics (Mad-CASP), a medium-high throughput approach that greatly improves proteome identification for single cells. The sample preparation of Mad-CASP can be performed in three steps in 3-4 hours, only requiring conventional lab supplies. Particularly, a synthetic hydrophobic peptide was coated on tubes to minimize sample loss during sample preparation and liquid chromatography (LC) separation. The peptide was rationally designed to be digested into smaller peptides of five amino acid residues and thus would not interfere with single-cell peptide identification during mass spectrometry (MS) acquisition and bioinformatics analysis. To cooperate with match-between-runs (MBR) function in MaxQuant, MS parameters were optimized by enhancing MS1 resolution to 240,000 and reducing fragment scan numbers in MS2 to Top 2. With Mad-CASP approach, peptide identification was largely improved, and 1240 ± 93 (n = 3) proteins were identified from a single HeLa cell. We further applied Mad-CASP to profile the proteome of 240 single human circulating CD34+ cells from both coronary chronic total occlusion (CTO) patients and healthy volunteers, discovering novel CD34+ subpopulations in CTO. Overall, this project introduced a pioneering single-cell proteomic approach with simplicity, convenience and effect, which is promising in large-scale proteome profiling.