Prostate cancer (PCa) exhibits significant intratumor heterogeneity, frequently manifesting as a multifocal disease. This study utilized Visium spatial transcriptomics (ST) to explore transcriptome patterns in PCa regions with varying Gleason scores (GSs). Principal component analysis (PCA) and Louvain clustering analysis revealed transcriptomic classifications aligned with the histology of different GSs. The increasing degree of tumor malignancy during GS progression was validated using inferred copy number variation (inferCNV) analysis. Diffusion pseudotime (DPT) and partition-based graph abstraction (PAGA) analyses predicted the developmental trajectories among distinct clusters. Differentially expressed gene (DEG) analysis through pairwise comparisons of various GSs identified genes associated with GS progression. Validation with The Cancer Genome Atlas Prostate Adenocarcinoma (TCGA-PRAD) dataset confirmed the differential expression of RBM39, a finding further supported by cytological and histological experiments. These findings enhance our understanding of GS evolution through spatial transcriptomics and highlight RBM39 as a gene associated with GS progression.