Objective To determine the role of serum cystatin C in evaluating the severity and predicting in-hospital mortality in patients with community-acquired pneumonia (CAP). Methods The clinical data of 176 patients with CAP treated between January 2015 and October 2016 were collected in a retrospective way. The CURB-65 score was used to assess the severity. The serum levels of cystatin C and C-reactive protein (CRP) on admission were measured. The correlations between cystatin C and CURB-65 score and between cystatin C and CRP were calculated. Receiver operating characteristic curve was used to determine the ability of cystatin C in predicting in-hospital mortality. Results The serum level of cystatin C increased with the increasing CURB-65 score (P<0.001). The serum level of cystatin C was correlated positively with CRP level (rs=0.190, P<0.011). There were 22 patients died in hospital, the mean serum cystatin C level of non-survivor was significantly higher than that of survivors [(1.51±0.56)vs. (1.02±0.29) mg/L, P<0.001]. At a cut-off 1.18 mg/L, the sensitivity and specificity of cystatin C in predicting in-hospital mortality were 68.18% and 81.17%, respectively. The area under the receiver operating characteristic curve was 0.793. The combination of cystatin C and CRP increased the predictive accuracy for in-hospital mortality. Conclusion Cystatin C level increases with the increaseing severity of CAP, and it may be a clinical biomarker to evaluate the severity and prognosis of patients with CAP.