BACKGROUND: Offering the overweight or obese patient the option of choosing from a selection of weight loss diets has not been investigated in type 2 diabetes. The aim of the study was to investigate if the option to choose from, and interchange between a selection of diets ("Choice"), as opposed to being prescribed one set diet ("No Choice"), improves drop out rates and leads to improved weight loss and cardio-metabolic outcomes. METHODS: The study was a 12 month, randomized parallel intervention. A total of 144 volunteers with type 2 diabetes or pre-diabetes and a BMI >27 were randomized to "No Choice" or "Choice". Those in the No Choice group were placed on a set weight loss diet (CSIRO) with no change permitted. Those in the Choice group could choose from, and interchange between, the CSIRO, South Beach or Mediterranean diets. RESULTS: There were no differences in attrition rates or weight loss between the "Choice" and "No Choice". In a secondary analysis of the intention-to-treat weight loss data with last measured weight carried forward gave a highly significant diet group by time by gender interaction (p = 0.002) with men doing better in the No Choice group overall (maximum difference "No Choice "-2.9 ± 4.6 kg vs. "Choice"-6.2 kg ± 5.3 kg at 6 months) and women doing better in the Choice group overall (maximum difference Choice -3.1 ± 3.7 kg vs. "No Choice" -2.0 kg ± 2.6 kg at 6 months). CONCLUSIONS: Men prefer direction in their weight loss advice and do less well with choice. A gender-specific approach is recommended when prescribing weight loss diets. TRIAL REGISTRATION: anzctr.org.au ACTRN12612000310864.