Minimal residual disease, or MRD, is an important prognostic indicator in childhood acute lymphoblastic leukemia. In ALL-IC-BFM 2002 study, we employed a standardized method of flow cytometry MRD monitoring for multiple centers internationally using uniformed gating, and determined the relevant MRD-based risk stratification strategies in our local patient cohort. We also evaluated a novel method of PCR MRD quantitation using peripheral blood plasma. For the bone marrow flow MRD study, patients could be stratified into 3 risk groups according to MRD level using a single time-point at day-15 (Model I) (I-A: <0.1%, I-B: 0.1-10%, I-C: >10%), or using two time-points at day-15 and day-33 (Model II) (II-A: day-15<10% and day-33<0.01%, II-B: day-15 ≥ 10% or day-33 ≥ 0.01% but not both, II-C: day-15 ≥ 10% and day-33 ≥ 0.01%), which showed significantly superior prediction of relapse (p = .00047 and <0.0001 respectively). Importantly, patients with good outcome (frequency: 56.0%, event-free survival: 90.1%) could be more accurately predicted by Model II. In peripheral blood plasma PCR MRD investigation, patients with day-15-MRD ≥ 10(-4) were at a significantly higher risk of relapse (p = 0.0117). By multivariate analysis, MRD results from both methods could independently predict patients' prognosis, with 20-35-fold increase in risk of relapse for flow MRD I-C and II-C respectively, and 5.8-fold for patients having plasma MRD of ≥ 10(-4). We confirmed that MRD detection by flow cytometry is useful for prognostic evaluation in our Chinese cohort of childhood ALL after treatment. Moreover, peripheral blood plasma DNA MRD can be an alternative where bone marrow specimen is unavailable and as a less invasive method, which allows close monitoring.