<p>Understanding the breeding potential of parental clones is essential for optimizing crossing strategies in sugarcane breeding. In this study, we evaluated the feasibility of introducing the best linear unbiased prediction (BLUP) methodology into the Japanese sugarcane breeding system by estimating breeding values (BVs) and specific combining ability (SCA) of parental clones and combinations for key traits. Initial estimations were based on datasets from four combining ability tests. The estimated BVs were consistent with field observations of progeny populations; for example, interspecific hybrids and their backcross generations exhibited higher BVs for stalk number but lower BVs for stalk diameter and Brix. Additionally, BVs for Brix were estimated using a historical seedling selection dataset, and a significant correlation was observed between these BVs and Brix data obtained from a recently conducted seedling selection. While a significant correlation was found between BVs estimated from the two datasets, the correlation of SCA values between them was not clear. Considering the broader range of BV variation and higher heritability observed in the seedling selection dataset, it is likely that estimations based on combining ability tests were limited by relatively narrow genetic diversity, whereas historical seedling selections captured a wider spectrum of genetic variation. These results suggest that the BLUP methodology can be effectively introduced and applied in Japanese breeding system. Furthermore, to fully leverage its potential, it is desirable to implement efficient phenotyping techniques for key traits within seedling selection trials, enabling the evaluation of a larger number of parental combinations.</p>