Objective To study the clinical characteristics in elderly inpatients with delirium and investigate the clinical application value of non-drug intervention. Methods A prospective trial was conducted on 200 elderly inpatients admitted to our department from June 2022 to June 2023. According to whether delirium occurred during hospitalization, they were divided into delirium group and non-delirium group. Clinical characteristics of delirium were statistically analyzed, and the value of non-drug intervention in improving their prognosis was studied. SPSS statistics 22.0 was used for statistical analysis. Intergroup comparison was performed using student′s t test or Chi-square test depending on data type. Multivariate logistic regression analysis was applied to identify the related factors affecting the occurrence of delirium in elderly inpatients. Results The incidence rate of delirium was 47.00% (94/200) in the 200 elderly inpatients, of which 30.85% (29/94) were accompanied by dementia. The delirium patients were mainly acute onset (85.11%, 80/94), and generally had clinical manifestations such as consciousness disorder, attention disorder, disorientation and cognitive dysfunction. Multivariate logistic regression analysis suggested that visual/auditory impairment (OR=5.562,95%CI 2.390-12.946), cognitive dysfunction (OR=3.823,95%CI 1.975-7.400), limited activity (OR=1.962,95%CI 1.287-2.990), malnutrition (OR=3.710,95%CI 1.638-8.401), anemia (OR=1.902,95%CI 1.638-8.401) and dementia (OR=2.210,95%CI 1.547-3.157) were risk factors for delirium in elderly inpatients (P<0.05). The risk prediction model of delirium in elderly inpatients was Logit (P)=-1.442+1.716×visual/auditory impairment+1.341×cognitive impairment+0.674×limited activity+1.311×malnutrition+0.643×anemia+0.793×dementia. Hosmer-Lemeshow test displayed that the model had goodness of fit, with Chi-square=21.157 and P<0.001, indicating that the model was significant in predicting delirium in elderly inpatients. Among the 94 elderly inpatients with delirium, 66 cases received only non-drug intervention, and the delirium duration was 1-14 d, with an average of (6.61±1.79) d. After one week of non-drug intervention, the score of Delirium Rating Scale-Revised-98 was significantly decreased than before intervention (P<0.05). Among the 66 patients with non-drug intervention, 40 (60.61%) were cured and 26 (39.39%) were improved at discharge. Conclusion The incidence of delirium is quite high in elderly inpatients. Most of them are acute onset, in hyperactive type, and with diverse and typical clinical manifestations. But dementia is more common in this group. Clinically, attention should be paid to the differentiation of dementia and delirium. The risk factors of delirium in elderly inpatients are mainly predisposing factors, and non-drug intervention shows generally better efficacy.