Background: It has been widely reported that the expression levels of SNHG20 are elevated in diverse types of cancers, indicating that SNHG20 may participate in cancer initiation and development. Besides, accumulating evidence reveals that SNHG20 overexpression is also connected with poor clinical outcomes among cancer patients. Herein, we carry out a systematic meta-analysis to further determine the prognostic and clinical significance of SNHG20 expression in various human cancers. Methods: Qualifying publications were selected by searching for keywords in PubMed, Embase, Web of Science and Cochrane Library databases, up to September 1, 2019. Pooled hazard ratio (HR) or odds ratio (OR) with corresponding 95% confidence interval (CI) was computed to estimate the strength of association between SNHG20 and survival of cancer patients or clinicopathology using Stata 14.0 software. Results: In total, 15 studies encompassing 1187 patients met the inclusion criteria were ultimately enrolled for analysis. According to the meta-analysis, patients with high SNHG20 expression were markedly linked to poorer overall survival (OS) (pooled HR = 2.47, 95% CI 2.05-2.98, P = 0.000) and disease-free survival/recurrence-free survival/progression-free survival (DFS/RFS/PFS) (pooled HR = 2.37, 95% CI 1.60-3.51, P = 0.000). Additionally, regarding clinicopathology of patients, enhanced SNHG20 was correlated with advanced tumour-node-metastasis (TNM) stage (OR = 2.80, 95% CI 2.00-3.93, P = 0.000), larger tumor size (OR = 3.08, 95% CI 2.11-4.51, P = 0.000), positive lymph nodes metastasis (OR = 2.99, 95% CI 2.08-4.31, P = 0.000), higher tumor stage (OR = 4.51, 95% CI 2.17-9.37, P = 0.000) and worse histological grade (OR = 1.95, 95% CI 1.44-2.63, P = 0.000), but not with gender, smoking status or distant metastasis. Conclusions: Up-regulated SNHG20 expression is ubiquitous in different kinds of cancers. Moreover, up-regulated SNHG20 expression is capable of serving as an innovative predictive factor of inferior clinical outcomes in cancer patients. Nevertheless, higher-quality multicenter studies are required to corroborate our results.