Introduction: The predictors of stent treatment failure and their importance 10 years after treatment with drug-eluting stents (DESs) have not been reported in detail. Methods: Data were retrieved from the SORT-OUT II database encompassing 2,849 non-left main coronary lesions in 2,073 unselected all-comer patients treated with first-generation DES and followed clinically for 10 years. Stent treatment failure (STF) was defined as definite or probable stent thrombosis, target lesion revascularization (TLR), or >70% restenosis left untreated. Target lesion failure (TLF) was defined as cardiac death, target vessel myocardial infarction, or TLR. Characteristics predicting higher hazard ratios (HRs) were identified by the multivariate Cox regression analysis. Results: A stent diameter ≤2.5 versus ≥3.5 mm had STF 23.3 versus 11.8% and TLF 27.9 versus 18.8%. Stent length <20 versus >40 mm had STF 13.0 versus 29.0% and TLF 18.7 versus 34.6%. In multivariate analysis, decreasing stent diameter (HR: 1.24 [3.0 mm] to 2.12 [2.25 mm], reference ≥3.5 mm) and increasing stent length (HR: 1.15 [20-30 mm] to 2.07 [>40 mm], reference <20 mm) predicted STF together with diabetes (HR: 1.31), previous revascularization (HR: 1.31), restenotic (HR: 2.25), bifurcation (HR: 1.45), and chronically occluded lesions (HR: 1.54). A predictive score (PS) was calculated for each lesion from the HRs for the predictors present. The 10-year rates of STF were 10% in lesions with a PS ≤ 1.5 and 37% in those with PS ≥ 3.5. Conclusions: Ten-year outcomes show large variations depending on the stent size and a few patient and lesion characteristics. The calculation of a PS from these unambiguous variables may be used to improve the risk estimate in individual lesions and patients.