Screening for early-stage disease is vital for reducing colorectal cancer (CRC)-related mortality. Methylation of circulating tumor DNA has been previously used for various types of cancer screening. A novel cell-free DNA (cfDNA) methylation-based model which can improve the early detection of CRC is warranted. For our study, we collected 313 tissue and 577 plasma samples from patients with CRC, advanced adenoma (AA), non-AA and healthy controls. After quality control, 187 tissue DNA samples (91 non-malignant tissue from CRC patients, 26 AA and 70 CRC) and 489 plasma cfDNA samples were selected for targeted DNA methylation sequencing. We further developed a cfDNA methylation model based on 11 methylation biomarkers for CRC detection in the training cohort (Area Under Curve (AUC) = 0.90 (0.85-0.94)) and verified the model in the validation cohort (AUC = 0.92 (0.88-0.96)). The cfDNA methylation model robustly detected patients pre-diagnosed with early-stage CRC (AUC = 0.90 (0.86-0.9)) or AA (AUC = 0.85 (0.78-0.91)). Here we established and validated a non-invasive cfDNA methylation model based on 11 DNA methylation biomarkers for the detection of early-stage CRC and AA. The utilization of the model in clinical practice may contribute to the early diagnosis of CRC.