Small cell carcinoma of the cervix (SCCC) is a rare primary neuroendocrine cervical carcinoma with a high degree of invasiveness. SCCC is prone to early-stage lymph node and distant metastases and characterized by a poor prognosis. Currently, there is no standard treatment. This study aimed to evaluate the clinicopathological factors and treatment models that influence SCCC prognosis through a systematic review and meta-analysis, to improve the diagnosis and treatment of SCCC. A comprehensive search was performed in multiple medical literature databases to retrieve studies on the clinical prognosis of SCCC published in China and abroad as of March 1, 2017. Twenty cohort studies with 1904 patients were analyzed. Meta-analysis showed statistical significance for the following factors: FIGO staging (hazard ratio [HR] = 2.63, 95% confidence interval [CI]: 2.13-3.24; odds ratio [OR] = 3.72, 95% CI: 2.46-5.62), tumor size (HR = 1.64, 95% CI: 1.25-2.15), parametrial involvement (HR = 2.40, 95% CI: 1.43-4.05), resection margin (HR = 4.09, 95% CI: 2.27-7.39), lymph node metastasis (OR = 2.09, 95% CI: 1.18-3.71), depth of stromal invasion (HR = 1.99, 95% CI: 1.33-2.97), neoadjuvant chemotherapy (HR = 2.06, 95% CI: 1.14-3.73), and adjuvant chemotherapy (HR = 1.63, 95% CI: 1.26-2.12; OR = 1.48, 95% CI: 1.02-2.16). FIGO staging, tumor size, parametrial involvement, resection margin, depth of stromal invasion, and lymph node metastasis can be used as clinicopathological characteristics for the prediction of SCCC prognosis. Neoadjuvant chemotherapy tended to improve prognosis. Our findings suggest that neoadjuvant chemotherapy plus adjuvant chemotherapy may be the preferred strategy. However, adjuvant radiotherapy appeared to cause no significant improvement in prognosis. Therefore, the clinical application of radiotherapy and the relationship between radiotherapy and clinicopathological factors need to be re-examined. The results of this study should be validated and developed in formal, well-designed multicenter clinical trials.