Maternal age is a risk factor for congenital heart disease even in the absence of any chromosomal abnormality in the newborn. Whether the basis of this risk resides with the mother or oocyte is unknown. The impact of maternal age on congenital heart disease can be modelled in mouse pups that harbour a mutation of the cardiac transcription factor gene Nkx2-5 (ref. 8). Here, reciprocal ovarian transplants between young and old mothers establish a maternal basis for the age-associated risk in mice. A high-fat diet does not accelerate the effect of maternal ageing, so hyperglycaemia and obesity do not simply explain the mechanism. The age-associated risk varies with the mother's strain background, making it a quantitative genetic trait. Most remarkably, voluntary exercise, whether begun by mothers at a young age or later in life, can mitigate the risk when they are older. Thus, even when the offspring carry a causal mutation, an intervention aimed at the mother can meaningfully reduce their risk of congenital heart disease.