The molecular landscapes of diffuse large B-cell lymphoma (DLBCL) remained to be comprehensively investigated with an urgent need to identify novel prognostic biomarkers guiding prognostic stratification and disease monitoring. Baseline tumor samples of 148 DLBCL patients were analyzed using targeted next-generation sequencing (NGS) for mutational profiling, whose clinical reports were retrospectively reviewed. In this cohort, the subgroup of old DLBCL patients (age at diagnosis > 60, N  = 80) exhibited significantly higher Eastern Cooperative Oncology Group scores and International Prognostic Index than their young counterparts (age at diagnosis ≤ 60, N  = 68). As revealed by the NGS results, PIM1 (43.9%), KMT2D (31.8%), MYD88 (29.7%), and CD79B (27.0%) were identified as the most frequently mutated genes. Aberrations of genes of the immune escape pathway were significantly enriched in the young subgroup, while the altered epigenetic regulators were more abundant in the old patients. FAT4 mutation was identified as a positive prognostic biomarker, associated with longer progression-free survival and overall survival in the entire cohort and the old subgroup, using the Cox regression analyses. However, the prognostic function of FAT4 was not reproduced in the young subgroup. We comprehensively analyzed the pathological and molecular characteristics of old and young DLBCL patients and demonstrated the prognostic value of FAT4 mutation, which requires further validation with sizable cohorts in future research.