Longan is an economically important fruit tree that is cultivated in subtropical regions worldwide. For longan breeding, traits associated with yield and quality are considered desirable. Thus, fine mapping of quantitative trait loci (QTLs) and mining candidate genes for longan fruit yield and quality would benefit longan breeding. In this study, we constructed the first single-nucleotide polymorphism (SNP)-based genetic map of an F1 longan population using restriction site-associated DNA sequencing. A total of 1,464,790,436 clean reads were generated from 200 F1 progeny from ‘Fengliduo’ (female) × ‘Dawuyuan’ (male) and their parents. The overall integrated map contained 8014 SNP markers, which were assigned to 15 linkage groups with a total length of 2873.39 cM, an average distance of 0.36 cM, and 99.96 % of gaps having a genetic distance of <5 cM. Based on two successive years of fruit phenotyping and this high-density linkage map, 17 stable QTLs were identified, including 12 for single fruit weight (SFW) and 5 for edible rate, but no stable QTLs were identified for soluble solid content. Finally, three key candidate genes, namely, Dlo_007693.1 (DlEXP4), Dlo_011045.1 (DlCNR8), and dlo_034894.1 (DlCYP450), for SFW were identified based on RNA sequencing data and quantitative reverse-transcription polymerase chain reaction. In conclusion, this is the first study on the construction of a high-density genetic map of longan using next-generation sequencing based on the largest mapping population. These results provide valuable information for marker-assisted selection breeding and mining of candidate genes related to fruit traits in longan.