Background Physical restraints are widely used in residential care. The objective was to investigate restraint use and its associations with neuropsychiatric symptoms and personal characteristics in residential care. Methods Data were collected in the South Savo Hospital District from 66 of 68 public or private institutions providing long‐term residential care. Nurses assessed the use of physical restraints and neuropsychiatric symptoms (NPS) included in the symptom list of the Neuropsychiatric Inventory (NPI). Drug use was obtained from medical records, and activities of daily living (ADL) were assessed by the nurses according to the Barthel Index. Results The total number of persons in residential care was 1386. Any restraint was used for 721 patients (52%) in the preceding 24 h. Bedrails were the most common restraints. In the multivariate analysis, psychotic symptoms (OR 1.94, 95% CI 1.14–3.31) and use of benzodiazepines (OR 1.69, 95% CI 1.18–2.41) were positively associated with restraint use, whereas antipsychotic (OR 0.62, 95% CI 0.44–0.87) and antidepressant drug use (OR 0.64, 95% CI 0.45–0.90) and higher ADL score (OR 0.9, 95% CI 0.92–0.93) were negatively associated. Concomitant use of at least two restraints was associated with high prevalence of hyperactivity NPS symptoms. Conclusions More than half of the residents were exposed to some physical restraint, most frequently bedrails, within the last 24 h. Psychotic symptoms and benzodiazepine use increased while good ADL and antipsychotic or antidepressant use decreased the risk of restraint use. Bedridden persons were the most frequently restrained which may pose an ethical problem. Copyright © 2016 John Wiley & Sons, Ltd.