The rice breeding process for grain yield could be effectively enhanced by developing efficient tools that accelerate plant selection through the rapid determination of reliable predictors. In this study, we have described various associations between grain yield and photosynthetic parameters, which can be easily and quickly obtained using a non-invasive technique on the flag leaf during the anthesis stage. Among the analyzed photosynthetic parameters, the photosynthetic performance index (PIABS) stood out due to its strong association with grain yield. A genome-wide association analysis conducted on plants from a rice diversity panel at the tillering stage revealed the presence of a quantitative trait locus on chromosome 9. This locus was characterized by a group of candidate chloroplastic genes that exhibited contrasting haplotypes for PIABS. An analysis of these haplotypes revealed a clear division into two groups. One group consisted of haplotypes linked to high values of PIABS, which were predominantly associated with Japonica spp. subpopulations. The other group consisted of haplotypes linked to low values of PIABS, which were exclusively associated with Indica spp. subpopulations. Japonica spp. genotypes exhibited higher values in the yield component panicle weight compared with the Indica spp. genotypes. The findings of this study indicate that PIABS could serve as an early predictor of yield parameters during the tillering stage in rice breeding processes.