Objectives. We aimed at evaluating both the efficacy and safety of TJ-54 (Yokukansan) in patients with treatment-resistant schizophrenia. This randomized, multicenter, double-blind, placebo-controlled study was conducted. Methods. One hundred and twenty antipsychotic-treated inpatients were included. Patients were randomized to adjuvant treatment with TJ-54 or placebo. During a 4-week follow-up, psychopathology was assessed using the Positive and Negative Syndrome Scale (PANSS). Results. TJ-54 showed a tendency of being superior to placebo in reduction total, positive, and general PANSS scores in treatment-resistant schizophrenia, but the difference was not statistically significant in both per-protocol set (PPS) and intention-to-treat (ITT). However, in PPS analysis, compared to the placebo group, the TJ-54 group showed statistically significant improvements in the individual PANSS subscale scores for lack of spontaneity and flow of conversation (TJ-54: -0.23 ± 0.08; placebo: -0.03 ± 0.08, P < 0.018), tension (TJ-54: -0.42 ± 0.09; placebo: -0.18 ± 0.09, P < 0.045), and poor impulse control (TJ-54: -0.39 ± 0.10; placebo: -0.07 ± 0.10, P < 0.037). Conclusions. The results of the present study indicate that TJ-54 showed a tendency of being superior to placebo in reduction PANSS scores in treatment-resistant schizophrenia, but the difference was not statistically significant. However, compared to the placebo group, TJ-54 group showed statistically significant improvements in the individual PANSS subscale scores.