Post-transplantation chimerism testing is important to monitor the engraftment of donor stem cells and for the diagnosis of relapse. Detecting the presence of donor/recipient-specific short tandem repeats (STRs) is a frequently used method for engraftment study. Unfortunately, the interpretation of the STR-based chimerism tests is often subject to interference by the presence of a stutter peak, which is one 4-base repeat unit smaller than an authentic allele. The aim of this study was to systematically analyze and resolve the effect of stutter peaks on the interpretation of STR-based chimerism tests. The AmpFlSTR Identifiler Amplification kit (Applied Biosystems)was used to amplify 15 STR loci using genomic DNA from 30 randomly selected, healthy donors. We found that the stutter peaks had locus-specific characteristics. The stutter percentage was defined as the percentage of the stutter peak area/main STR peak area. Based on mean values for the 30 DNA samples, the stutter percentage varied from locus to locus and ranged from 3.12% to 10.71% for 15 STR loci. The locus-specific stutter effect can be eliminated through appropriately adjusted equations. The usefulness of these equations in the prediction of relapse was confirmed by the 5% sensitivity test. Hence, this report offers a valuable scheme to enhance the accuracy of chimerism testing.