Genome structure may influence the accuracy of genomic prediction, but the effect of chromosome number under different levels of heritability remains insufficiently understood. This study evaluated the effects of chromosome number and heritability on genomic prediction accuracy using a controlled simulation framework. Four populations with 5, 10, 20, and 30 chromosomes were simulated, with each chromosome containing 1,000 markers and 50 QTL. Three heritability levels (0.10, 0.30, and 0.50) were considered, resulting in 12 simulation scenarios with 20 independent replicates per scenario. Genomic breeding values were predicted using GBLUP, and prediction accuracy was assessed as the Pearson correlation between genomic estimated breeding values and true breeding values. A two-way factorial ANOVA was used to evaluate the effects of chromosome number, heritability, and their interaction. Both chromosome number and heritability significantly affected prediction accuracy (P < 0.001), whereas their interaction was not significant (P = 0.956). Mean prediction accuracy increased from 0.328 at heritability 0.10 to 0.497 and 0.587 at heritability 0.30 and 0.50, respectively. The five-chromosome population showed significantly higher accuracy (0.512) than the 10-, 20-, and 30-chromosome populations (0.468, 0.448, and 0.453, respectively), while these three populations did not differ significantly. These results demonstrate that both trait heritability and genome-scale characteristics can affect GBLUP prediction accuracy and should be considered when designing and interpreting genomic prediction simulations.