<p>The optimal strategy for patients with in-stent restenosis (ISR) is controversial. We aimed to compare the effects of a drug-coating balloon (DCB) and drug-eluting stent (DES) in ISR treatment.<strong> </strong>Clinical trials were extensively collected, and retrieved items were screened for inclusion. Both clinical (major adverse cardiac event (MACE); myocardial infarction (MI); and target lesion revascularization (TLR) and angiographic (minimal lumen diameter (MLD), and stenosis relative to reference lumen diameter (SRLD) endpoints were extracted and compared. MACE and MI were not significantly different between the groups. Pooled results of TLR showed a marginal effect that DES was superior to DCB (13.50% for DCB vs<em>.</em> 11.17% for DES, RR = 1.256, 95% CI: 0.997 to 1.583, <em>P</em> = 0.053), with heterogeneity across studies (<em>I</em><sup>2</sup> = 42.0%, Cochrane <em>Q</em>-test = 0.069). Meta-regression identified bare metal stent (BMS) or drug eluting stents (DES) implanted in the previous intervention and proportions of diabetes in the DCB group as sources of heterogeneity. DES implantation also significantly improved angiographic outcomes (WMD for MLD: −0.318, 95% CI: −0.424 to −0.213, <em>P</em><em> </em>&lt; 0.001; WMD for SRLD: 6.164%, 95% CI: 4.915% to 7.412%, <em>P</em><em> </em>&lt; 0.001). All DES, including everolimus-eluting ones, did not benefit BMS-ISR patients compared with DCB treatment.<strong> </strong>DES implantation, which is superior to DCB angioplasty only in DES-ISR patients, should be preferred in the DES-ISR population to reduce TLR. DCB may be preferred in BMS-ISR to avoid increasing stent layers.</p>