This longitudinal study using the sampling frame of the second Australian prevalence study of psychosis aimed to identify predictors of the health and well-being of care-givers of people with psychosis and inform social work recovery-oriented practice. Ninety-eight carers were recruited at baseline and seventy-eight re-interviewed after one year. Correlational and regression analyses were conducted to identify relationships between carer and service user-related factors and predictors of carers’ health and well-being over time. Carers’ poor health and well-being were predicted by a combination of specific service user-related characteristics and care-giving factors. Carers’ assessment of the functioning of their relative/friend with psychosis at baseline had stronger relationships with their own health and well-being at follow-up than other factors. Carers’ care-giving burden was predicted by their educational levels and their relative/friend’s cognitive levels over time. To achieve improved health and well-being for carers, services need to consider potential deterioration of carers’ physical health over time, to facilitate appropriate referral of carers with physical health problems and provide psycho-social rehabilitation services to improve the functioning of people with psychosis. Findings provide some evidence to support social work recovery-oriented practice in working with people with psychosis and the routine inclusion of carers in such interventions.