Pneumonia is an important infectious entity that affects residents in long-term care facilities (LTCFs), whereas hospitalization-requiring pneumonia (HRP) represents a more critical patient condition with worse outcomes. The evidence addressing the association between Barthel index and risk of HRP among LTCF residents is lacking. A multicenter, retrospective cohort study was conducted in three LTCFs enrolling adult patients who resided for 3 months or more and ever underwent Barthel index evaluation within a study period of January 1 to December 31, 2010. The endpoint was HRP after enrollment. A total of 299 patients (169 women; age, 79.0 ± 12.2 years) were enrolled and categorized into HRP Group (n = 68; 36 women; age, 79.1 ± 11.3 years) and Non-HRP Group (n = 231; 133 women; age, 79.0 ± 12.4 years) by the endpoint. The patients in HRP Group had significantly lower Barthel index (8.6 versus 25.8 points, p < 0.001) but higher proportion of chronic obstructive pulmonary disease (13.2% versus 3.9%, p = 0.004). By the multivariate analysis of logistic regression, we found that lower Barthel index (odds ratio (OR), 0.967; p < 0.001), existence of chronic obstructive pulmonary disease (OR, 4.192; p = 0.015), and feeding route (percutaneous endoscopic gastrostomy comparing with oral feeding; OR, 0.177; p = 0.012) were independently associated with HRP. In conclusion, a lower Barthel index is significantly associated with the occurrence of pneumonia that requires hospitalization in long-term care residents. Barthel index is a useful and reliable tool for risk evaluation in this population.