Abstract: OBJECTIVE To investigate the risk factors of pulmonary infection, and evaluate the value of combined diagnosis of brain natriuretic peptide（BNP）, tumor necrosis factor-α（TNF-α） and lipoprotein（a）（Lp（a）） in the diagnosis of the disease. METHODS From May 2016 to Aug. 2019, 45 elderly stroke patients with lung infection in Fuyang District First People’s Hospital were recruited as the infection group, and 51 elderly stroke patients without lung infection were included as the uninfected group. The clinical characteristics of the two groups were statistically compared, and the factors affecting the occurrence of stroke and lung infection in the elderly were explored. The correlation between BNP, TNF-α, Lp（a） levels and neurological function（NIHSS score） in elderly stroke patients with pulmonary infection and their value in diagnosing elderly stroke with pulmonary infection were analyzed. RESULTS The occurrence of elderly stroke with lung infection was affected by the Glasgow Coma Scale（GCS） score, NIHSS score, long-term smoking history, combined with underlying diseases, accompanied by swallowing disorders, the number of invasive operations ≥3 times, the use of antibacterial drugs ≥2, and the time spent in bed for more than 7 days, stroke type and serum BNP, TNF-α, Lp（a） levels（P＜0.05）. There was a positive correlation between serum TNF-α, BNP, Lp（a） levels and NIHSS score in elderly patients with stroke and lung infection（P＜0.05）. Among the three serum indexes, the area under the curve（AUC） of serum TNF-α in the diagnosis of elderly stroke with pulmonary infection was the highest at 0.818, with diagnostic sensitivity of 77.78% and specificity of 80.39%. CONCLUSION The occurrence of senile stroke with pulmonary infection is affected by many factors such as the degree of brain injury, the degree of neurological deficits, the increase of serum BNP, TNF-α, and Lp（a） levels. Among them, serum BNP, TNF-α, Lp（a） levels can reflect the neurological deficits in the infected patients, and may have high application value in the early diagnosis of stroke with pulmonary infection.