We conducted propensity-score-matched comparisons of midterm angiographic follow-up outcomes of sirolimus- versus everolimus-eluting stents (SES, EES) after elective placements for de novo coronary stenosis in small vessels (SV) in patients with diabetes mellitus (DM), because the angiographic efficacy of EES over SES for those cohorts remained unclear. The study was a non-randomized, retrospective, lesion-based, multicenter study, examining lesions followed up angiographically within 550 days, extracted from the unified database of 6 institutes. The endpoint (binary restenosis) was defined as the percentage of subjects having >50% diameter stenosis at follow-up. Propensity-score-matched analyses were conducted in 3 different vessel-size cohorts, defined by a preprocedural reference diameter (RD) <2.10, <2.35, and <2.60 mm, yielding group sizes of n = 107, 183, and 312 baseline-adjusted lesions in each of the 2 stent arms. The frequency of binary restenosis decreased significantly with increasing vessel size, at 16.8, 12.6, and 12.2%, in the SES group. However, it remained almost the same across vessel-size groups in the EES group (8.0, 6.0, and 7.5%). The p values for the significance of the differences in binary restenosis between EES and SES in each vessel size increased with the decrease in vessel size [p = 0.002, 0.040, and 0.063 (the last still nearly significant)]. Thus, in patients with DM, EES showed increasingly superior efficacy over SES for SV stenosis as the vessel size became smaller, i.e., the risk for binary restenosis became greater.