Abstract Background T-cell lymphoma (TCL) is highly aggressive and has a poor prognosis; thus, it is worth exploring biomarkers that may predict clinical outcomes and investigate their potential role in developing targeted therapies. In this study, we characterized the mutation pattern of tumor necrosis factor-alpha-inducing protein 3 (TNFAIP3) and its role in the prognosis of TCL patients. Methods Coding sequence (CDS) mutations in TNFAIP3 in TCL patients was explored using exome-sequencing data from 79 patients in our center (Guangdong Provincial People’s Hospital, GDPH) and 544 samples from the Catalogue of Somatic Mutations in Cancer (COSMIC) database. Additionally, non-CDS mutations in TNFAIP3 in 41 TCL patients from our center (JNU) were investigated by polymerase chain reaction (PCR) and Sanger sequencing. Furthermore, non-CDS mutations in TNFAIP3 in 47 TCL patients from Gene Expression Omnibus (GEO) dataset were explored. Results In the COSMIC database, TNFAIP3 mutations in TCL patients were located in the CDS, and the overall mutation frequency was 2.2%. However, TNFAIP3 mutations were not detected in the CDS of any of the samples in our center’s datasets. Interestingly, non-CDS TNFAIP3 mutations were found in 14.6% and 4.3% of TCL patients in the JNU and GSE15842 dataset, respectively. Importantly, there was a clear trend showing that TCL patients with a TNFAIP3 mutation were associated with a longer 5-year restricted mean survival time (RMST) and favorable OS rate compared with those without a TNFAIP3 mutation in the JNU dataset [hazard ratio (HR) = 0.29, 95% confidence interval (CI) 0.07 to 1.31, P = 0.089]. Furthermore, TNFAIP3 mutations significantly correlated with T-cell large granular lymphocytic leukemia (T-LGLL) with a favorable prognosis in the JNU dataset (P = 0.002). Notably, the different mutation patterns of TNFAIP3 when comparing our center and the COSMIC datasets might be due to different ethnic and genetic backgrounds. Conclusions To the best of our knowledge, we for the first time describe that TNFAIP3 mutations in non-CDS regions are associated with favorable OS for TCL patients, which might be a potential biomarker for the prognostic stratification of Chinese TCL patients.