Background Presupplementary motor area (pre‐SMA) hyperactivity has been detected in obsessive–compulsive disorder (OCD) patients. However, it is not understood whether this is a putative primary cause or a compensatory mechanism in OCD pathophysiology. Considering the polarity‐dependent effects on cortical excitability of transcranial direct current stimulation (tDCS), we applied cathodal and/or anodal tDCS to the pre‐SMA of OCD patients to test which current polarity might better improve symptoms. Methods Twelve OCD patients received initially 10 anodal (n = 6) or cathodal (n = 6) daily consecutive 2 mA/20 min tDCS sessions with the active electrode placed bilaterally on the pre‐SMA. In case of improvement or no change in symptoms severity, the subjects were maintained on the same current polarity for 10 more sessions. In case of symptoms worsening after the first 10 sessions they were switched to the other polarity for 10 more sessions to test the hypothesis of a polarity‐dependent effect. Therefore, each subject received 20 tDCS sessions. The Yale‐Brown Obsessive–Compulsive Scale (Y‐BOCS) and the Sheehan Disability Scale (SDS) were administered biweekly to assess changes in symptoms severity. Results After 10 sessions, 50% of patients who initially received anodal stimulation were switched to cathodal, while 100% of patients initially assigned to cathodal stimulation continued on the same polarity. At the end of the study, a statistically significant decrease was observed in the mean Y‐BOCS scores of those patients who underwent cathodal tDCS. No pre–post difference was found in the scores of patients following anodal tDCS. Conclusions Cathodal but not anodal tDCS over the pre‐SMA significantly improved OCD symptoms.