Photodynamic diagnosis (PDD)-assisted transurethral resection of bladder tumors (TURBT) has emerged as a promising complementary tool to white light (WL) cystoscopy, potentially improving cancer detection and replacing conventional mapping biopsies. This study aimed to investigate the diagnostic accuracy of PDD by anatomical locations in mapping biopsies through lesion-based analysis. PDD and WL findings were prospectively recorded in 102 patients undergoing mapping biopsies and PDD-assisted TURBT using oral 5-aminolevulinic acid. We evaluated 673 specimens collected from flat tumor or normal-looking lesions on WL cystoscopy, after excluding 98 specimens collected from papillary or nodular tumors. Among the 673 lesions, cancer was detected in 110 (16%) by lesion-based analysis. PDD demonstrated significantly higher sensitivity (65.5% vs. 46.4%, p < 0.001) and negative predictive value (92.5% vs. 89.5%, p < 0.001) compared to WL. The sensitivity of PDD findings varied by location: posterior (100%), right (78.6%), dome (73.3%), left (70.6%), trigone (58.8%), bladder neck (41.7%), anterior (40.0%), and prostatic urethra (25.0%). Incorporating targeted biopsies of specific locations (bladder neck, anterior, and prostatic urethra) into the PDD-guided biopsies, regardless of PDD findings, significantly increased the overall sensitivity from 65.5% to 82.7% (p = 0.001). This study first demonstrated the detection rate of location-specific mapping biopsies using PDD, revealing difficulties in accuracy assessment in areas susceptible to tangential fluorescence. While PDD-guided biopsy improves cancer detection compared to WL cystoscopy even for flat tumors or normal-looking lesions, more careful decisions, including mapping biopsies, may be beneficial for an assessment in these tangential areas.