Simultaneous improvement of grain yield, milling quality, and eating quality requires coordinated selection across correlated traits and assessment of whether line performance remains consistent across successive generations. Two segregating japonica rice populations, P1 (Jinchuan No. 1 × Jinyujing 22; 74 lines) and P2 (Koshihikari × Jinyujing 22; 45 lines), were evaluated in the F₄ and F₅ generations for grain yield and indicators of milling and eating quality. Multiple analytical approaches covering intergenerational trait variation, correlation analysis, radar charts, principal component analysis and two-dimensional comprehensive-consistency evaluation were used. Several eating-quality-related indicators showed significant correlations between the F₄ and F₅ generation-year observations, whereas grain yield showed no significant F₄–F₅ correlation in either population. Five P1 lines and eight P2 lines combined above-average composite performance with relatively high apparent F₄–F₅ intergenerational consistency. These lines represent candidates for subsequent validation in replicated multi-environment trials, and the results illustrate the potential value of jointly considering composite performance and apparent intergenerational consistency in early-generation rice selection.

UAV multispectral imagery can support rice yield estimation, but spectral–yield relationships may change with cultivar and planting density. We evaluated three japonica rice cultivars under five planting-density treatments in a split-plot experiment and collected UAV multispectral imagery on 13 dates. We examined five vegetation indices: NDVI, RVI, NDRE, OSAVI, and CLRE. Cultivar significantly affected grain yield and yield components. Planting density had no significant main effect, while the cultivar × density interaction was significant for grain yield. Spectral–yield correlations were strongest during 19–25 August, which broadly matched reproductive development based on field observations. NDRE and CLRE showed the most consistent associations. A four-index PLSR model using 19 August data achieved an outer-LOOCV R² of 0.773 and an RMSE of 530.28 kg ha⁻¹. Tuned RF gave comparable performance (R² = 0.783, RMSE = 518.60 kg ha⁻¹). Adjustment for cultivar, planting density, and block substantially reduced the pooled mid-August correlations. However, the four-index spectral set still improved prediction beyond cultivar, planting density, and their interaction. Grouped validation changed little when density levels or cultivar × density combinations were withheld, but performance declined sharply when an entire cultivar was excluded. These results show that mid-August UAV observations were informative in this experiment, but wider use will require validation across more cultivars, years, and sites.