Objective To analyze the risk factors and construct prediction model for occult pneumonia (OP) in patients with senile dementia. Methods Medical records of patients with confirmed diagnosis of senile dementia complicated with pulmonary infection and treated in the Third Affiliated Hospital of Anhui Medical University from Ja-nuary 2019 to December 2020 were retrospectively analyzed. Some patients were randomly selected from the confirmed patients as the modeling group, they were divided into OP group, non-OP group, and other cases were as the verification group. Univariate and logistic regression multivariate analysis were used to analyze the risk factors for OP in patients with senile dementia, nomogram was constructed by R4.0.3 software and the model was verified. Results A total of 216 patients were included, 148 of whom (75 cases of OP and 73 cases of non-OP) were used for modeling, and 68 cases (37 cases of OP and 31 cases of non-OP) were used for verification. Diabetes mellitus (OR=2.565, 95%CI: 1.094-6.015), severe dementia (OR=3.079, 95%CI: 1.116-8.494), dementia duration ≥ 10 years (OR=5.782, 95%CI: 2.139-15.627), age ≥ 80 years (OR=2.737, 95%CI: 1.011-7.413), and long-term bedridden (OR=4.835, 95%CI: 1.716-13.625) were independent risk factors for senile dementia complica-ted with OP (all P < 0.05). The prediction model was constructed and verified through the five risk factors, the verification results showed that the area under the curve (AUC) in modeling group was 0.841 and that in verification group was 0.756, suggesting that the diagnostic ability of the model was good; Hosmer-Lemeshow test showed that the model had goodness of fit, decision curve analysis showed that this model had high benefit. Conclusion Age ≥ 80 years, severe dementia; dementia duration ≥ 10 years; diabetes mellitus, and long-term bedridden are independent risk factors for OP in patients with senile dementia, the individualization of nomogram model can predict the probability of OP in patients with senile dementia, so as to intervene as soon as possible and improve the prognosis.