BACKGROUND Nursing is considered as a risk profession with high levels of stress and burnout, and these levels are probably increasing. OBJECTIVE A model of prediction of burnout in nursing that includes socio-demographic variables, job stressors, and personal vulnerability, or resistance, is proposed. METHODS A cross-sectional correlational design was used. A sample of 473 nurses and student nurses in practice from three General Hospitals in Madrid (Spain) completed the "Nursing Burnout Scale". The data were analysed using descriptive statistics, Pearson correlations, and hierarchical multiple regression. RESULTS The proposed model is a good predictor of the diverse burnout sub-dimensions: emotional exhaustion, depersonalisation, and lack of personal accomplishment. Significant predictors of burnout included age, job status, job stressors (workload, experience with pain and death, conflictive interaction, and role ambiguity), and hardy personality (commitment, control, and challenge). CONCLUSIONS Identifying an integrative process of burnout among nurses is an essential step to develop effective managerial strategies so as to reduce the burnout problem. Specifically, the present study suggests that intervention aimed at reducing the risk for burnout may achieve better results if it includes enhancement of workers' hardy personality rather than just decreasing environmental stressors.