Muscle fiber types, numbers and area are crucial aspects associated with meat production and quality. However, there are few studies of pig muscle fibre traits in terms of the detection power, false discovery rate and confidence interval precision of whole-genome quantitative trait loci (QTL). We had previously performed genome scanning for muscle fibre traits using 183 microsatellites and detected 8 significant QTLs in a White Duroc × Erhualian F2 population. The confidence intervals of these QTLs ranged between 11 and 127 cM, which contained hundreds of genes and hampered the identification of QTLs. A whole-genome sequence imputation of the population was used for fine mapping in this study.A whole-genome sequences association study was performed in the F2 population. Genotyping was performed for 1,020 individuals (19 F0, 68 F1 and 933 F2). The whole-genome variants were imputed and 21,624,800 SNPs were identified and examined for associations to 11 longissimus dorsi muscle fiber traits.A total of 3201 significant SNPs comprising 7 novel QTLs showing associations with the relative area of fiber type I (I_RA), the fiber number per square centimeter (FN) and the total fiber number (TFN). Moreover, one QTL on pig chromosome 14 was found to affect both FN and TFN. Furthermore, four plausible candidate genes associated with FN (KNDC1), TFN ( KNDC1) and I_RA (SLC36A4, CNTNAP5 and GRM8) were identified.An efficient and powerful imputation-based association approach was utilized to identify genes potentially associated with muscle fiber traits. These identified genes and SNPs could be explored to improve meat production and quality via marker-assisted selection in pigs.