This study aims to use integrated bioinformatics technology to dig a predictive miRNA-signature correlated with the prognosis of Nasopharyngeal carcinoma (NPC). Initially, 94 up-regulated and 91 down-regulated differentially expressed microRNAs (DEMs) of NPC were identified in the GEO dataset, and univariate COX regression analysis showed their abnormal expression were significantly associated with poor prognosis of NPC, respectively (P=0.002, P<0.001 and P<0.001). Subsequently, hsa-miR-29c, hsa-miR-30e and hsa-miR-93 identified by random forest algorithm were used to construct a predictive signature through multivariate COX regression analysis. Moreover, PCA, Kaplan-Meier analysis, time-dependent ROC analysis, and univariate and multivariate COX regression analysis demonstrated that there were significant differences in risk score, survival time and the expression of 3 DEMs between the high-risk group and the low-risk group (P< 0.001), and the high-risk group had worse OS (P<0.001). Furthermore, the average AUC values of 1- to 5-year OS, DFS and DMFS predicted by the signature were all above 0.7, and showed better clinical independence than other commonly used indexes. Eventually, 295 differentially expressed mRNAs were obtained by the intersection of the predicted results of TargetScan and the differentially expressed mRNAs in the datasets. Among them, 56 differentially expressed mRNAs were related to PFS. GO and KEGG enrichment analysis indicated that the poor prognosis of NPC was related to the abnormality of chromosomes, cytokines, and chemokines. Collectively, we constructed a 3-miRNA signature with better independent performance in predicting the prognosis for NPC, which may provide a basis for finding new therapeutic targets of NPC patients.