There is an absence of standardized validated instruments to assess the complex needs of pregnant women and mothers with severe mental illness. We aimed to develop a standardized assessment of need for pregnant women and mothers with severe mental illness. Staff and service users were asked to identify relevant domains of need. Professional experts and service users were then surveyed and asked to rate the importance of the domains of the Camberwell Assessment of Need - Mothers version (CAN-M). Reliability was established using 36 service user-staff pairs. Concurrent validity was assessed with the Global Assessment of Functioning. Inter-rater reliability (concordance) coefficients for unmet needs were 0.93 (95% confidence interval 0.89 to 0.98) (service users) and 0.83 (95% confidence interval 0.73 to 0.94) (staff); test-retest reliability coefficients were 0.91 (95% confidence interval 0.86 to 0.97) and 0.85 (95% confidence interval 0.73 to 0.96), respectively. Relevant CAN-M domains correlated with the Global Assessment of Functioning-symptom (Spearman's r correlation coefficient = -0.36, 95% confidence interval = -0.62 to -0.04, p = 0.05) and Global Assessment of Functioning-disability subscales (Spearman's r correlation coefficient = -0.52, confidence interval = -0.73 to -0.23, p < 0.01). We conclude that the CAN-M is a reliable and valid instrument for assessing the needs of pregnant women and mothers with severe mental illness.