Objective To establish a model for prognosis analysis of severe community-acquired pneumonia in order to find the independent risk factors for mortality. Methods The data of 88 patients with severe community-acquired pneumonia enrolled from 533 community-acquired pneumonia patients in Fujian Provincial Hospital from April 2012 to December 2015 were analyzed, they were divided into a survival group and a death group according to prognosis. The clinical materials of basic data of the population, clinical manifestation, treatment and prognosis and pulmonary severity indexes were collected. Then univariate analysis was used to screen risk factors of death before logistic multivaritae regression was applied to explore independent risk factors. Results The different pathogen groups including viral, bacterial, mixed infection, negative and other groups were compared and no differences were found in mortality (all P>0.05). Univariate analysis revealed antibiotics treatment before admission, higher APACHEⅡ score, higher Chalison's score, septicopyemia, and higher level of procalcitonin, blood urea nitrogen (BUN), blood glucose, lactate could increase death risk for the patients. While antiviral treatment and no invasive mechanical ventilation were determined as protective factors. Logistic multivaritae regression showed three factors including higher lactate and higher serum BUN and higher heart rates were independent death risk factors [OR values were 4.704 (95%CI 0.966-22.907), 1.264 (95%CI 0.994-1.606), and 1.081 (95%CI 1.003-1.165), respectively]. Whereas no invasive mechanical ventilation was protective factor (OR=0.033, 95%CI 0.001-0.764). Conclusion The patients with higher lactate and BUN, higher heart rate and accepting invasive mechanical ventilation have poor prognosis.