<p>Drought is a critical stress threatening global rice (<i>Oryza sativa</i>) production, with yield losses often exceeding 25%. Leveraging genetic engineering to develop drought-resilient rice through heterologous expression of drought-tolerant genes from diverse species represents a promising solution. Although the role of synonymous codon usage bias (CUB) in modulating gene expression has been recognized for decades, its impact on the expression of drought-responsive genes at different developmental stages in rice remains poorly understood. This study investigates CUB in drought-responsive genes during three key developmental stages: tillering, booting, and panicle elongation. Our results show a clear link between the variation in codon usage and the GC content of these genes, highlighting its impact on gene expression under drought conditions. High-expression genes with elevated GC content at the third codon position exhibit stronger codon bias compared to low-expression genes. Notably, drought conditions shift the preferred codons for the key amino acids, including for proline, serine, threonine, histidine, asparagine and aspartic acid, relative to basal codon usage in rice. Analysis of neutrality plot and Ka/Ks ratio indicates that purifying selection predominantly drive codon usage, favouring codons that optimize translational efficiency. These insights establish a basis for rational gene design to enhance heterologous expression and development of drought-tolerant rice varieties through genetic engineering.</p>