Introduction People with severe mental illness (SMI) and learning disabilities (LD) experience significant health inequalities, with diet-related conditions contributing substantially to early and preventable death. Despite high levels of nutritional risk, the presence and effectiveness of nutritional screening in mental health (MH) and LD settings remains under-researched. This study aimed to investigate nutritional screening practices in UK inpatient MH and LD services from the perspectives of dietitians. Methods A cross-sectional mixed-methods study was conducted using a novel 22-question online survey. Data was collected via the British Dietetic Association Mental Health Specialist Group (April-June 2025). Quantitative data was analysed descriptively and qualitative data by reflexive thematic analysis. Findings were integrated and presented thematically. Ethical approval was granted by Teesside University (2025Mar26544). Results Forty-seven dietitians participated, most with substantial dietetic experience, from a range of MH settings. Screening practices were widely established and supported by policy and audit. However, participants reported low confidence in screening translating into meaningful patient care. Barriers to screening included appropriateness of available tools, time constraints, difficulty engaging distressed patients and poor prioritisation of physical health. Digital integration and wider infrastructure were also important. Dietitians rarely undertook screening directly, instead holding secondary or leadership roles, while screening was most often completed by nursing staff who were often perceived to place limited importance on the process. Existing tools, particularly the Malnutrition Universal Screening Tool (MUST), were viewed as insufficiently capturing the broader nutritional risks relevant to MH/LD populations, leading some services to adopt bespoke, unvalidated tools. Conclusion Concerns regarding the suitability of existing nutritional screening tools in MH/LD settings are consistent with previous literature. However, we suggest cautious use of unvalidated bespoke tools. Whilst there was no clear front runner, MH specific tools such as the St Andrews Nutrition Screening Instrument (SANSI) and the NutriMental Screener warrant further evaluation. Importantly, findings indicate that optimising tool choice alone is unlikely to improve screening effectiveness. Nutritional screening must be embedded within clear care pathways, supported by organisational leadership, digital infrastructure, and multiprofessional engagement to move beyond procedural completion and support meaningful clinical action to improve patient care.