Amylopectin branch-chain length distribution is a key determinant of rice starch functionality and eating quality, yet the genetic basis underlying specific chain fractions remains incompletely understood in diverse germplasm. In this study, amylopectin fine structure was quantified in 137 accessions of the Korean Rice Core Selection (KRICE_Core) using HPAEC-PAD, and genome-wide association studies were conducted with 2.1 million high-quality SNPs under the FarmCPU model. The short-chain distribution (SCD, DP 6-12) varied from 26.85% to 37.20%, whereas the intermediate-chain distribution (ICD, DP 12-24) ranged from 52.57% to 61.04%, and the two fractions showed a strong inverse correlation. GWAS identified two major loci on chromosomes 4 and 6, with the chromosome 6 region showing exceptionally strong association signals and co-localizing with Os06g0229800 (SSIIa). Haplotype analysis of SSIIa detected five non-synonymous SNPs defining five haplotypes. Hap1 and Hap5 were associated with increased SCD and reduced ICD, whereas indica-related haplotypes (Hap2-Hap4) showed the opposite trend, indicating that SSIIa allelic variation quantitatively shifts the balance between short and intermediate chain fractions. In a panel of 113 Korean-bred cultivars, Hap1 and Hap5 together accounted for > 99% of genotypes, suggesting strong directional selection for japonica-type SSIIa alleles during Korean breeding. These findings provide robust molecular targets for designing rice varieties with tailored starch architecture and cooking qualities.