Aim: To compare the outcomes of GDPT [gemcitabine (G), cisplatin (D), prednisone (P), thalidomide (T)] versus CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) in treating newly diagnosed PTCL (peripheral T-cell lymphoma). Methods: An open-label prospective clinical trial with 153 newly diagnosed PTCL patients conducted between January 2010 and December 2018 was designed. Patients were randomly assigned to the GDPT (77 cases) and CHOP (76 cases) groups. Patients in each group were further divided into four subgroups: PTCL, not otherwise specified (PTCL-NOS); anaplastic large cell lymphoma (ALCL), angioimmunoblastic T cell lymphoma (AITL), and other types subgroup, in accordance with pathological patterns. Based on expression of RRM1, TOP2A, TUBB3, and ERCC1, patients were divided into groups with high and low gene expression levels. Clinical characteristics, side effects, efficacy, progression-free survival (PFS), and overall survival (OS) were compared. Results: There were no significant differences in the basic clinical features or side effects between the GDPT and CHOP groups. The overall response rate (ORR) of the GDPT group was better than that of the CHOP group (66.3% versus 50.0%, p = 0.042), as was the complete remission (CR) rate (42.9% versus 27.6%, p = 0.049). Patients in the GDPT group had a longer PFS and OS than the CHOP group. The 4-year PFS and OS rates in the GDPT group were both superior to those in the CHOP group (63.6% versus 53.0% for PFS, p = 0.035; 66.8% versus 53.6% for OS, p = 0.039). In the GDPT group, the difference in CR between the four subgroups was statistically significant ( p = 0.046). In the CHOP group, differences in both CR and ORR among the four subgroups were statistically significant ( p < 0.001 and p = 0.005, respectively). There were also statistically significant differences in CR between patients treated with CHOP and GDPT in the PTCL-NOS subgroup, AITL subgroup, and the other types subgroup ( p = 0.015; p = 0.003; p = 0.005, respectively). The data also showed a significant difference in OS among the four subgroups within the GDPT group ( p = 0.001). The OS of AITL was shorter than that of the other three subgroups. Four subgroups of CHOP showed a significant difference in PFS ( p = 0.019). There was no statistical association between responses and the gene expression levels of RRM1, ERCC1, TUBB3, and TOP2A. Conclusion: The GDPT group had better response rates and prolonged patient PFS and OS. As a promising new regimen, GDPT is expected to become the first-line therapy for PTCL. New agents should be applied to patients who do not achieve good responses with previous treatment, such as those diagnosed with angioimmunoblastic T cell lymphoma. Trial registration: This open randomized prospective clinical trial was registered at ClinicalTrials.gov (NCT01664975).