Objective To explore the mediation effects of nutrition and cognition in Apolipoprotein E Allele 4 (APOE ε4) with neuropsychiatric symptoms (NPSs) in Alzheimer′s Disease (AD). Methods A spectrum of 293 AD patients from amnestic mild cognitive impairment (aMCI) to AD dementia stage was consecutively collected from the Alzheimer′s Disease Biomarkers and Lifestyle Study (CIBL) cohort of Beijing Tiantan Hospital Affiliated to Capital Medical University from June 2021 to January 2023. The patients were divided into the carrier group (with APOE ε4, n=107) and the non-carrier group (without APOE ε4, n=186). Patients in significant NPS sub-groups were further analyzed in age, gender, body mass index (BMI), mini-mental state examination (MMSE), Montreal cognitive assessment (MoCA), and micro-nutritional assessment (MNA). SPSS 26.0 was used for data processing. According to the data type, t test, Mann-Whitney U test or χ² test was used for comparison between groups, and statistically significant factors in the hypothesis test were determined as confounding factors in the subsequent mediating effect analysis. A simple mediation effect model was used to analyze the potential mediating effects of nutritional status and cognitive function on APOE ε4 and NPS and its subtypes. Results Compared with non-carrier group, the carrier group had higher rates of hallucinations, apathy, and abnormal motor behaviors, the differences were statistically significant (P<0.05). The NPS subtype patients were analyzed separately. Compared with non-hallucination group, patients in hallucination group were older, had lower education years, carried a higher proportion of APOE ε4, and lower MMSE, MoCA and MNA scores. Compared with non-apathy group,patients in apathy group had a higher proportion of males, carried a higher proportion of APOE ε4, and had lower MMSE, MoCA and MNA scores.Compared with non-aberrant motor activity group, patients in aberrant motor activity group were older, had a lower proportion with hyperlipidaemia, carried a lower proportion of APOE ε4, and had lower MMSE, MoCA and MNA scores, with statistically significant differences (P<0.05). After adjusting for confounding factors, MNA scores mediated 29.80% (95%CI 0.062-0.522) of the correlation between APOE ε4 and apathy and 19.95% (95%CI 0.011-0.419) of the correlation between APOE ε4 and aberrant motor behaviors. The correction mediated by MMSE scores on APOE ε4 and hallucinations, apathy, aberrant motor behaviors were 24.21% (95%CI 0.078-0.573), 39.01% (95%CI 0.155-0.914), and 23.37% (95%CI 0.068-0.576). Conclusion For patients with aMCI and AD, nutritional status and cognitive function partially mediate the correlation between APOE ε4 and apathy and aberrant motor behaviors, and cognition also partially mediate the correlation between APOE ε4 and hallucination.