Background: High-quality evidence on morale in the mental health workforce is lacking. Aims: To describe staff well-being and satisfaction in a multicentre UK National Health Service (NHS) sample and explore associated factors. Method: A questionnaire-based survey (n = 2258) was conducted in 100 wards and 36 community teams in England. Measures included a set of frequently used indicators of staff morale, and measures of perceived job characteristics based on Karasek's demand-control-support model. Results: Staff well-being and job satisfaction were fairly good on most indicators, but emotional exhaustion was high among acute general ward and community mental health team (CMHT) staff and among social workers. Most morale indicators were moderately but significantly intercorrelated. Principal components analysis yielded two components, one appearing to reflect emotional strain, the other positive engagement with work. In multilevel regression analyses factors associated with greater emotional strain included working in a CMHT or psychiatric intensive care unit (PICU), high job demands, low autonomy, limited support from managers and colleagues, age under 45 years and junior grade. Greater positive engagement was associated with high job demands, autonomy and support from managers and colleagues, Black or Asian ethnic group, being a psychiatrist or service manager and shorter length of service. Conclusions: Potential foci for interventions to increase morale include CMHTs, PICUs and general acute wards. The explanatory value of the demand-support-control model was confirmed, but job characteristics did not fully explain differences in morale indicators across service types and professions.