Background: Concurrent chemoradiotherapy (CCR) with cisplatin plus 5-Fu (PF) regimen is a standard modality for locally advanced esophageal squamous cell carcinoma (ESCC) patients. This trial aimed to assess the efficacy and safety of the paclitaxel plus 5-Fu (TF) regimen versus PF regimen in CCR for ESCC patients. Methods: ESCC patients presenting with stage IIa to IVa were enrolled in a prospective multicenter phase 3 study. Patients were randomized to either TF or PF group. Patients in TF group were treated with 5 cycles of weekly TF (5-Fu 300 mg/m², civ 96h plus paclitaxel 50 mg/m², d1) in CCR followed by 2 cycles of monthly TF (5-FU 1800 mg/m², civ 72h, plus paclitaxel 175 mg/m² d1) in consolidation chemotherapy. Patients in PF group were treated with 2 cycles of CCR followed by 2 cycles of consolidation chemotherapy with PF (cisplatin 25 mg/m²/d, d1-3, plus 5-Fu 1800 mg/m², civ 72h, q28d). The radiotherapy dose was 61.2 Gy delivered in 34 fractions. The primary end-point was the 3-yr OS. Results: 436 ESCC patients (217 assigned to TF group and 219 assigned to PF group) in 6 centers were recruited between April 2012 and July 2015. Median follow-up of patients who survived was 44.6 months [IQR 29.3–72.0]. The 3-yr OS was 57% in TF group and 51% in PF group (HR 0.91; 95% CI 0.69-1.18; P = 0.46). No significant differences were recorded in 3-yr DPFS or 3-yr LPFS between TF and PF groups (44.3% vs. 45.3% and 48.8% vs. 49.8%, respectively). TF group had a significant higher incidence of acute Grade 3/4 leukopenia (31.3% vs. 18.3%), dermatitis (5.1% vs. 1.4%), and pneumonitis (9.7% vs. 3.2%), and significant lower incidence of anemia (0.5% vs. 3.2%), thrombocytopenia (0.5% vs. 13.7%), fatigue (6.9% vs. 19.6%), anorexia (1.4% vs. 14.6%), nausea (1.4% vs. 14.2%), and vomiting (2.3% vs. 18.3%) than PF group (P < 0.05). There were 3 (1.4%) patients in TF group died of acute pneumonitis, 1 (0.5%) patient in TF group and 2 (0.9%) patients in PF group died of delayed pneumonitis. Conclusions: TF might be an option used in CCR in ESCC patients with a different type of side effects compared with PF, although it did not significantly prolong OS. Clinical trial information: NCT01591135