Male pattern baldness is the most common type of progressive hair loss, which has a considerable negative impact on personal image and mental health. Recently, several large sized genome-wide association studies (GWASs) in European and American populations have reported a large number of MPB-associated genetic susceptibility sites, but the genetic effects of these loci in East Asian populations remain unclear. This study was based on a sample of 684 Eurasian males in China, and conducted a population heterogeneity analysis on 624 MPB-related single nucleotide polymorphisms (SNPs) found in a large GWAS consisting of 205,327 European males in the UK Biobank. Polygenic risk scores (PRS) were used to build predictive models, and the relationship between the number of predictive factors and the predictive performance of the model was investigated. A total of 467 SNPs passed quality control, among which 6.9% showed nominally significant association with MPB (P <0.05). Linear and logistic models established by including the effects of age, body mass index (BMI), and 25 SNPs demonstrated fairly accurate prediction accuracy (R_² = 28.9%, AUC = 0.82). Age had the greatest impact on the model performance (R_² = 22.9%, AUC = 0.77), and the AUC reached the maximum when BMI and 68 SNPs were added cumulatively (about 0.89). This study indicates that MPB has strong genetic heterogeneity in European and East Asian populations. Selecting a subset of SNPs can achieve prediction accuracy close to that of the complete set. The prediction model constructed here may help understand the genetic mechanism, early diagnosis and prevention of MPB in East Asian populations.