Objective: To compare the efficacy and tolerability of olanzapines and haloperidol in treating patients with amphetamine psychosis. Material and Method: Fifty-eight patients experiencing episode of amphetamine psychosis were randomly assigned to olanzapine (N=29) or haloperidol (N=29) in 1:1 (olanzapine: haloperidol) ratio. All patients started with 5-10 mg/day of the study drug; after each 7-day period, the study drug could be adjusted in 5-mg increments or decrements within the allowed dose range of 5-20 mg/day during the 4-week double-blind period. Results: Clinical response was seen in both treatment groups since the first week. Ninety three percent of the olanzapine patients (N=27 of 29) and 79.3% of the haloperidol patients (N=23 of 27) were clinically improved at endpoint. These differences were not statistically significant (p=0.25). The Simpson-Angus total score change from baseline to endpoint reflected no extrapyramidal symptoms among the olanzapine-treated patients (median=0.0, range=0.0). In contrast, worsening occurred among the haloperidol-treated patients (median=0.2, range=0.0-3.1). The differences of mean change in Simpson Angus Scale significantly favored olanzapine (p<0.01). Change to endpoint on the Barnes Akathisia Scale showed that olanzapine-treated patientsû scores were close to the baseline (median=0.0, range=-1.0-0.0), whereas haloperidol-treated patientsû scores worsened from the baseline (median=0.0, range=-1.0-3.0). This difference was statistically significant (p=0.02). Conclusion: Both olanzapine and haloperidol were efficacious in the treatment of patients with amphetamine psychosis. Olanzapine was superior to conventional neuroleptic haloperidol in treatment safety with lower frequency and severity of extrapyramidal symptoms.