Gestational diabetes mellitus (GDM) is a type of diabetes which develops, or is first recognised during pregnancy. GDM increases the risk of medical complications during pregnancy, and its consequences extend to later in life, as both women and their children are at high risk of developing cardio-metabolic disease. Obesity and excessive gestational weight gain (GWG) are modifiable risk factors for GDM development. Therefore, the research in this thesis aimed to develop and evaluate dietary interventions to help prevent GDM. I initially explored the full burden of the disease using data from a big UK cohort. I found that women with GDM are at high risk of premature death. GDM increases the risk of subsequently developing cardiovascular disease (CVD) risk factors (diabetes, hypertension, dyslipidaemia), and increases the risk of incident CVD by 50% compared to no GDM. There was no evidence that the CVD risk factors explained the association of GDM with CVD, suggesting that all women with GDM are at increased risk of cardio-metabolic disease post-partum, and should be monitored. I then developed a dietary intervention of moderate total carbohydrate reduction, suitable for use in pregnancy, to help prevent GDM (RECORD). It combined a 30-minute dietary behavioural consultation, with structured written materials, brief telephone support throught pregnancy, and weekly self-weighing. In designing the behavioural component, I sought ways to increase motivation and adherence. I thought about using Motivational Interviewing (MI), a popular counselling style in behavioural weight management programmes (BWMPs). Due to increased training and time required for its delivery, I conducted a systematic review and meta-analysis to explore MI’s independent effectiveness over other behavioural approaches in BWMPs for weight loss or maintenance, but found that it only leads to a small short-term weight benefit (-1 kg), and only when compared to no/minimal interventions, which makes it not a worthwile addition to weight management programmes with standard behavioural support. In a feasibility trial, I tested if it is possible to deliver the RECORD intervention alongside routine antenatal appointments, and how successful women with obesity would be in following it from early or mid-pregnancy until delivery. The trial’s retention rate was excellent (96%), but recruitment was difficult. Adherence was highly variable, and on average, the reduction in total carbohydrate intake in women receiving the intervention was small (-19.5 g/day and -24.6 g/day absolute and adjusted changes from baseline), which limits the likelihood of achieving the desired goal to prevent GDM. I observed potentially favourable effects of the intervention on blood glucose, GWG, and blood pressure, but the trial was not powered to detect differences in these outcomes. In qualitative interviews, participants praised the intervention for its structure and simplicity, and generally found it acceptable, but a sub-sample reported barriers to adherence, such as competing priorities, and emotional relationship with food. Combining qualitative and quantitative data from this research, I identified opportunities for refinement of the intervention, recruitment strategy, and for routine practice, which could be explored in the future.