There is evidence for a deleterious effect of maternal mental illness on mother-infant interaction. Presence of mental illness and lowered maternal cognitive function independently predict quality of interaction, but their combined effect on interaction is unclear. A pilot study was conducted to explore the relationship between maternal serious mental illness (SMI), cognitive function, and mother-infant interaction. Six mothers with SMI (two with schizophrenia and four with depression) and 12 mothers with no psychiatric history were recruited, together with their infants (all aged between S and 18 weeks). Mother-infant interaction was assessed using the Crittenden Care Index. The Cognitive Drug Research computerised assessment battery provided measures of memory and attention. In support of previous literature, mothers with SMI were significantly less sensitive during interactions with their infants. They also demonstrated impairment relative to controls on an index of speed of memory processing. A hierarchical regression model revealed that presence of maternal mental illness was a significant predictor of maternal sensitivity, but when speed of memory processing was accounted for, the strength of this relationship was reduced, suggesting that the relationship between maternal mental illness and mother-infant interaction may be partially mediated by level of cognitive function. Further study is warranted.