AIM: The mobility trajectories of hospitalized older people are heterogeneous, and there is a need to understand patient groups with similar trajectories. Our aims were to identify mobility trajectories of older patients admitted to geriatric wards in an English university hospital, and describe patient characteristics and outcomes associated with different mobility groups. METHODS: A retrospective observational study of first admissions to geriatric wards for 3 months starting 1 May 2016 was carried out. Anonymized routinely collected clinical data was extracted from the electronic patient records: age, sex, Clinical Frailty Scale score, presence of delirium (yes/no), diagnosis of dementia (yes/no), Charlson Comorbidity Index, falls in the past year and the highest recorded Modified Early Warning Score in the Emergency Department. Physiotherapists routinely collected the Elderly Mobility Scale on admission and discharge; this information was entered in a K-means cluster analysis to identify mobility trajectory groups. We compared clinical characteristics and outcomes between these groups and used logistic regression models to determine independent predictors of group membership. RESULTS: The overall functional trajectory was improvement, from a median Elderly Mobility Scale of 11 on admission to 14 on discharge. Cluster analysis identified five mobility trajectory groups: low-low (n = 117), intermediate-intermediate (n = 95), intermediate-high (n = 205), low-high (n = 89) and high-high (n = 209). Higher frailty and the presence of cognitive impairment were associated with the low-low and intermediate-intermediate group membership, and negatively associated with the high-high group membership. CONCLUSIONS: The majority of hospitalized patients experienced functional improvement. Frailty and cognitive impairment might be associated with poorer mobility trajectories and hence require early identification. Geriatr Gerontol Int 2019; 19: 305-310.