This study compared the anti-aggressiveness effects of the atypical anti-psychotic olanzapine with that of selective serotonin reuptake inhibitors (SSRI) and benzodiazepines (BZD) among patients with heroin dependence submitted to opioid-agonists substitution treatment. Sixty-seven (67) patients who met the DSM-IV criteria for heroin dependence and showed aggressive personality traits, not affected by comorbid schizophrenia or bipolar disorder, accepted to participate in a 12-week prospective, observational trial. Patients were included into two subgroups in relationship with treatment, for the evaluation of the endpoints at week 12: group 1: substitution treatment in combination with OLA (32 patients); group 2: substitution treatment in combination with fluoxetine/paroxetine and clonazepam (35 patients). Efficacy measures were Buss Durkee Hostility Inventory (BDHI), Symptoms Check List-90 (SCL 90) anger–hostility scores, incidence rates of aggressive incidents and attacks. The rates of patients who remained in treatment at week 12 in group 1, treated with OLA, and group 2, treated with SSRI and BDZ, were not significantly different (17 = 53.1% vs 16 = 45.7%). BDHI total, direct aggressiveness, verbal aggressiveness scores, SCL 90 aggressiveness scores and aggressive incidents rates showed a significantly more consistent decrease from baseline in group 1 than in group 2 subjects, in the patients who completed the treatment ( p < 0.001; p < 0.01; p < 0.05; p < 0.01; p < 0.001). Among the completers, 69.3% achieved early full substance abuse remission, while 30.7% achieved partial substance abuse remission, with no significant difference between 1 and 2 treatment subgroups. Although obtained by an observational–open clinical study, with multiple limitations, our findings suggest that OLA may be useful as an adjunctive agent in reducing aggressive/hostile behaviour in heroin addicted individuals during maintenance substitution treatment. Otherwise, atypical anti-psychotic OLA seems to be unable to improve the outcome in terms of addictive behavior and relapse risk in the addicted patients not affected by overt psychotic disorders.