Aims: Salivary duct carcinoma (SDC) is a highly aggressive salivary gland malignancy. The UICC/AJCC 9th edition (TNM-9th) introduced a revised N-classification based on metastatic lymph node counts. However, prior studies have reported inconsistent findings regarding the prognostic value of size-based T classification. This study aimed to validate the prognostic performance of the TNM-9th in SDC and to explore a refined staging model that reflects its biological characteristics. Methodology: This retrospective multi-institutional cohort study included 517 patients with centrally confirmed SDC who underwent curative-intent surgery at 11 institutions between 1994 and 2023. Overall survival (OS) and disease-free survival (DFS) were evaluated. Prognostic performance of the UICC/AJCC 8th (TNM-8th) and TNM-9th was compared with that of a refined staging system model using Harrell’s concordance index (C-index) and the Akaike Information Criterion (AIC). Results: Among the 517 patients, the nodal count–based N-classification in TNM-9th improved prognostic stratification compared with the TNM-8th. In contrast, the conventional size-based T-classification showed limited prognostic discrimination, with no significant differences in OS or DFS between T1 and T3 categories in either edition. The refined staging system, which merges T1–T3 to avoid size-based upstaging, demonstrated improved predictive accuracy for OS (C-index 0.676; AIC 2472.6) compared with TNM-8th (0.631; 2504.0) and TNM-9th (0.663; 2480.1). In addition, under the TNM-8th or TNM-9th, 18.7% of non-invasive carcinoma ex pleomorphic adenoma cases were classified as Stage I, whereas this proportion was 86.2% under the refined system. Conclusion: Although the nodal count–based N-classification in TNM-9th improves prognostic stratification, the conventional size-based T-classification may be inadequate for SDC. The proposed staging system, emphasizing invasive extent and nodal burden rather than tumor size, provides more accurate risk stratification.