Background: Pneumocytis pneumonia (PCP) is a life-threatening disease in non-HIV infected children. However, there have been few studies that have examined the clinical characteristics associated with PCP and outcomes for these pediatric patients. Objectives: A retrospective review was performed over a 10-year period to evaluate the clinical characteristics and outcome of non-HIV children diagnosed with PCP at Beijing Children's Hospital in China. Results: A total of 60 non-HIV children diagnosed with PCP were included in the study. The overall mortality was 41.7% (25/60). Underlying diseases included connective tissue disease (n = 23; 38.3%), hematological disease (n = 14; 23.3%), nephrotic disease (n = 8; 13.3%) and immunodeficiency disease (n = 10; 16.7%). In all, 26/40 (65.0%) children developed PCP after receiving a follow-up large dose of glucocorticoid because of recurrent disease. Median time from beginning glucocorticoid medication to PCP diagnosis was 245.9 days (range: 14-2100 days). The area under the ROC curve of CD4/CD8 T cell levels for the diagnosis of PCP was 0.902 (95% confidence interval, 0.849-0.955). The analysis rendered an optimum cut-off value of 0.715 corresponding to 89.2% sensitivity and 80.4% specificity. Using a multivariate logistic regression model, three parameters were identified as significantly associated with mortality: LDH level, mechanical ventilation and co-infection. Conclusion: The outcome of PCP in non-HIV children remains poor. A critical stage for PCP development is administration of follow-up glucocorticoid without prophylaxis. CD4/CD8 ratio is a suitable biomarker for predicting PCP and diagnostic of PCP in non-HIV children. Poor prognostic factors include LDH level, need for mechanical ventilation and co-infection.