Abstract Objective To explore the efficacy and safety of 5% lidocaine‐medicated plaster (LMP) in patients with trigeminal neuralgia (TN). Background TN is an excruciatingly painful type of neuropathic facial pain. Anti‐epileptics are the first‐line treatment for TN; however, these oral drugs alone sometimes fail to achieve satisfactory analgesic effects. Two retrospective studies have shown that LMP can be an effective and safe treatment option for some patients with TN. No other high‐quality clinical studies have explored the effect and safety of LMP in patients with TN. Methods The PATCH trial is an enriched enrollment with randomized withdrawal, double‐blind, vehicle‐controlled, parallel‐group trial performed at five study centers. Eligible patients with TN received LMP during a 3‐week initial open‐label phase. Patients who met the response criteria entered the double‐blind treatment phase and were randomly assigned for treatment with either LMP (LMP group) or vehicle patches (control group) at a 1:1 ratio. Patients who met the criteria for treatment failure were withdrawn from the double‐blind treatment phase, and treatment was continued in the remaining patients for up to 28 days. The primary outcome was the number of treatment failures. The secondary endpoints were the time to loss of therapeutic response (LTR) in the double‐blind phase and the weekly mean pain severity in both the open‐label phase and the double‐blind phase of the study. Results The first patient was enrolled in this study on May 1, 2021, and the enrollment of the last patient was completed on August 26, 2022. A total of 307 patients were initially screened, 226 (74.0%) of whom entered the open‐label phase. Of the 226 respondents, 124 (55.0%) were randomized to the double‐blind phase. In the double‐blind phase, 62 patients were assigned to the LMP group, and 62 were assigned to the control group. For the primary endpoint, 16 (26.0%) patients with LMP and 36 (58.0%) patients with vehicle patches met the treatment failure criteria during the double‐blind phase (relative risk, 0.48; 95% confidence interval [CI], 0.31 to 0.75; p < 0.001). The survival curve of the LTR showed that the LTR of LMP was significantly longer than that of the vehicle patches (hazard ratio, 0.275; 95% CI, 0.15 to 0.50; log‐rank p < 0.001). LMP also significantly reduced the weekly mean pain severity in the double‐blind phase of the study ( p = 0.007). Conclusions LMP produced partial relief of pain symptoms in some patients with TN. For responders, LMP may be used as an add‐on therapy in a multidrug treatment protocol.