Background: TROP2 (TACSTD2) is a therapeutic target of antibody–drug conjugates and is broadly expressed in epithelial malignancies. Data on TROP2 protein expression in large, clinically annotated prostate cancer cohorts, particularly in recurrent disease, remain limited. Methods: TROP2 immunohistochemistry was performed on a prostate cancer tissue microarray comprising 17,747 primary radical prostatectomy specimens and 258 recurrent prostate cancer samples. TROP2 staining data were compared with clinicopathological parameters, biochemical recurrence and ERG status. Results: Among 12,807 interpretable primary prostate cancers, TROP2 staining was detectable in all cases and considered strong in 94.5%, moderate in 3.5%, and weak in 2.0%. Reduced TROP2 was statistically associated with high pT stage (p = 0.0011) and high Gleason grade (p < 0.0001), but absolute differences were small. Reduced TROP2 expression levels were unrelated to PSA recurrence in the entire cohort (p = 0.0787) but statistically linked to early PSA recurrence in ERG positive cancers (p = 0.0165). In four different scenarios of multivariate analyses, reduced TROP2 expression did not show an unequivocal independent prognostic role. As compared to primary tumors, TROP2 was markedly reduced among 250 evaluable recurrent cancers (strong 57.6%, moderate 31.6%, weak 8.4%, negative 2.4%; p < 0.0001 vs. primary cancers). Reduced TROP2 was also significantly more frequent in recurrent cancers than in the subgroup of high-grade primary tumors (Gleason ≥ 4+4; strong 57.6% vs. 90.2%; p < 0.0001). Conclusions: TROP2 is consistently expressed at high levels in primary prostate cancer but often reduced in recurrent disease. While uniform expression in primary tumors supports TROP2 as a therapeutic target in prostate cancer, reduced expression in many recurrent cancers highlights the potential need for TROP2 expression measurement as a predictive test in this subgroup.