Background: Effective risk assessment techniques and follow-up strategies are scarce for patients with prostate cancer because of the heterogeneity of the population. We aimed to conduct a study for better risk stratification, applying machine learning–based classification to patients with prostate cancer who underwent major surgery. Methods: We identified eligible patients who underwent prostatectomy or surgeries related to excision of the prostate using data from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) for the years 2016–2021. Hierarchical clustering was conducted with variables such as age, BMI, smoking status, taking HTN medication, past medical history (diabetes, COPD, and CHF), preoperative labs (Na, BUN, creatinine, WBC count, HCT, and platelet count), operation time in minutes, and major complications. A multivariate logistic regression analysis was also conducted with a 30-day readmission as an outcome. The model accounted for cluster groups (high risk vs. low risk), among other variables. Results: The median [Q1, Q3] age of the sample (n = 83,776) was 65.0 [59.0, 69.0] years. The participants were divided into two clusters (low-risk cluster, n = 72891; high-risk cluster, n = 10885) by hierarchical clustering, using 31 clinical variables. Patients in the high-risk cluster showed higher median levels of BUN (21 vs. 16 mg/dL), WBC count (7 vs. 6.5 /uL), platelet count (225 vs. 221/μL), and operation time (231 vs. 197 minutes). 30-day readmission rates were higher among the high-risk cluster (12.1% vs. 5.4%; p< 0.001). Similar trends were observed for minor complications (10.7% vs. 5.2%), major complications (8% vs. 2.5%), and reoperation (3.4% vs. 1.3%). Results from logistic regression showed that cases in the high-risk cluster were 1.10 (95% CI: 1.03, 1.17; p< 0.0015) times as likely to return to the hospital within 30 days of discharge as compared to the cases in the low-risk cluster (AUROC: 78.6%). Conclusions: In our study, hierarchical clustering divided prostate cancer cases into subgroups with a higher 30-day readmission risk. Applying machine learning–derived classification to prostate cancer patients undergoing major surgery may contribute to better risk stratification, leading to better prognosis and follow-up strategies.