Objective:To investigate determinants of antipsychotic medication use among older people living in aged care homes in Australia. Design: Retrospective study of a random sample of de-identified medication reports using cross-sectional data gathered between 1 January 2008 and 30 June 2008 in Australia. Subjects: The mean (SD) age of the residents was 84.0 (9.0) years. Seventy-five per cent were females. Measures: Resident demographics, clinical characteristics, medical diagnoses and prescribed medication were systematically recorded. Logistic regression (LR) models were used to determine predictors for any antipsychotic, atypical and conventional antipsychotic use. Results: Twenty-three per cent of the residents were prescribed one or more antipsychotics. In the LR model, factors for predicting the odds ratio and 95 confidence interval (CI) for any antipsychotic medication use were agitation (7.11, 95 CI 3.15-16.03), challenging behaviours (7.47, 95 CI 2.53- 22.10), dementia (2.35, 95 CI 1.36-4.06), dementia with mood disorder (0.39, 95 CI 0.16-0.92), paranoia (6.70, 95 CI 1.08-41.55), psychosis (14.79, 95 CI 3.64-60.00) and any psychiatric diagnosis (3.30, 95 CI 1.82-6.00). Use of atypical antipsychotic medication was significant for agitation (4.58, 95 CI 2.05-10.23), aggression (2.25, 95 CI 1.05-4.78), challenging behaviours (8.01, 95 CI 2.76- 23.24), dementia (3.64, 95 CI 1.99-6.67), dementia with mood disorder (0.16, 95 CI 0.06-0.43), psychosis (16.51, 95 CI 4.28-63.66) and any psychiatric diagnosis (4.44, 95 CI 2.33-8.46). Conclusions: Psychiatric diagnosis, psychosis and dementia were associated with significantly greater odds for the use of antipsychotic medications. Older people suffering from dementia and comorbid mood disorders treated with antidepressants were less likely to be prescribed atypical antipsychotics.