Abstract: 【Objective】 Based on genome-wide association analysis，the relevant genes affecting the milk production and milk quality of buffaloes were screened out，and the potential functions of these genes were verified，which could provide theoretical basis for accelerating the breeding of high-quality and high-yield buffalo breeds. 【Method】 Seventy buffaloes（36 crossbred buffaloes，18 Murrah buffaloes，and 16 Nili-Ravi buffaloes）were selected for the measurement of phenotypic values related to milk production traits. Genomic DNA was extracted from these 70 buffaloes and subjected to reduced genome sequencing. Principal component analysis and genome-wide association studies were conducted in conjunction with the milk production traits of buffaloes. Based on the identified SNP loci，candidate genes were searched， followed by GO functional annotation analysis and KEGG signaling pathway enrichment analysis. 【Result】 All 70 buffaloes had complete phenotypic values，and there were extremely significant correlations among these milk production traits （P<0.01）. The 70 buffaloes did not show obvious population stratification and could essentially be regarded as a population. A total of 9 candidate genes were identified near 12 SNPs associated with buffalo milk production，among which 3 genes，CDH8，SLC30A7 and FANCC，were key genes related to the milk production of buffaloes；28 candidate genes were identified near 32 SNPs associated with buffalo peak milk production，among which 12 genes，Yeats4，SMC6， SUMO1，Grin2a，INO80，GEN1，Exd1，GDPD4，ABCG5，LRRK2，Arx and GRM5，were key genes related to the peak milk production of buffaloes；1 candidate gene（MID1）was identified near 3 SNPs associated with buffalo milk protein rate，which was significantly enriched in GO functional terms and KEGG signaling pathways（P<0.05，the same below）； 1 candidate gene（CNTN5）was identified near 3 SNPs associated with buffalo milk fat rate，which was significantly enriched in GO functional terms and KEGG signaling pathways；1 candidate gene（FAM160A1）was identified near 4 SNPs associated with buffalo milk protein content，but this gene was not enriched in any GO functional terms or KEGG signaling pathways；and 6 candidate genes were identified near 7 SNPs associated with buffalo milk fat content，and CTNNA2 ，SSBP2 and IPOll genes were key gene related to buffalo milk fat content. 【Conclusion】 Based on genome-wide association analysis，46 candidate genes for milk production traits are identified from 70 buffaloes，among which 20 are key genes for milk production traits in buffaloes. Therefore，these genes can be selected as new molecular marker resources in production to accelerate the breeding of superior buffalo breeds. Furthermore，based on genome-wide association analysis，genes related to milk production in buffaloes can be efficiently mined，opening up a new direction for quantitative trait research.