Severely energy-restricted diets (SERDs) are an effective treatment for obesity, however, adherence to such diets is often perceived as poor by healthcare professionals. This investigation evaluated adherence to a 12-week SERD in participants with class II and III obesity. Reported food consumption was compared against individualised SERD prescriptions. Body weight measures were obtained at baseline, 12 and 52 weeks. The data were analysed in three groups (i) the entire cohort (n = 26), (ii) completers (n = 13) and (iii) non-completers (n = 13). SERD prescription elements included (i) the number of meal replacement products; (ii) total protein; (iii) total energy intake; (iv) level of dietary energy restriction; (v) vegetable serves; (vi) water serves, and (vii) how much physical activity was performed. A generalised repeated-measures mixed-effects model was used to investigate if adherence to the program elements individually, or collectively, influenced weight loss. Completers had an average (± SD) of 4549 ± 748 kJ energy intake per day, resulting in a mean energy restriction of 62% compared to the 69% prescribed, indicating a degree of non-adherence. The percent weight changes for completers and non-completers were −7.8 ± 4.7% and −1.6 ± 2.6% at 12 weeks, and −12.2 ± 12.1% and −1.8 ± 3.2% at 52 weeks, respectively. Complete dietary adherence to a SERD may not be necessary to achieve a clinically relevant weight loss of 12% at 52 weeks, if energy is restricted by at least 62% (~4600 kJ per day) relative to requirements.