Associations between daily functional trajectories and new-onset all-cause dementia and Alzheimer's disease (AD) and the role of body weight are underexplored. Data were from the Health and Retirement Study (HRS) 1994–2020. Daily function was assessed using (instrumental) activities of daily living ([I]ADLs). All-cause dementia and AD were defined by self- or proxy-reported physician diagnoses. Body weight was assessed using body mass index (BMI) and categorized as normal (18.5 kg/m2 ≤ BMI < 30 kg/m2) and abnormal (BMI < 18.5 kg/m2 or ≥30 kg/m2). The group-based trajectory modeling and Cox proportional hazards regression were utilized. Of 18,763 adults included, 1236 developed new-onset dementia during a 10-year follow-up. The associations of ADL and IADL limitations at baseline with all-cause dementia and AD were much more pronounced in those with abnormal weight (P for interaction < 0.005). Five joint trajectories of ADL and IADL limitations were identified: No (72.7 %), Recovery (4.0 %), Recent emerging (16.4 %), Early emerging (4.8 %), and Severe (2.1 %). Furthermore, the 'Severe' joint trajectory (vs. 'No') was associated with 3.57- and 3.59-times higher risks of new-onset all-cause dementia and AD in participants with abnormal weight (P for interaction = 0.002 and 0.005). Notably, the Recovery joint trajectory (vs. No) was not associated with increased risks of all-cause dementia or AD. Self-/proxy-reported all-cause dementia and AD may introduce misclassification bias. Lifestyle factors were not quantified. BMI at baseline, but not its trajectory, was utilized. Potential reverse causation deserved attention. Body weight control can help reduce the risk of progression from functional limitations to all-cause dementia and AD.