Recently, patients who received induction chemotherapy plus concurrent chemoradiotherapy for locoregionally advanced nasopharyngeal carcinoma were found to have survival advantages compared with those receiving concurrent chemoradiotherapy alone in two large randomized trials. Based on these two trials, we present a cost-effectiveness analysis to compare gemcitabine and cisplatin (GP) versus cisplatin, fluorouracil, and docetaxel (TPF) for induction chemotherapy to treat locoregionally advanced nasopharyngeal carcinoma. We constructed a Markov model to compare the cost and effectiveness of GP versus TPF. Clinical data including the frequency of adverse events, recurrence and death obtained from two randomized phase III trials were used to calculate transition probabilities and costs. Health utilities were estimated from the literature. Incremental cost-effectiveness ratios, expressed as dollars per quality-adjusted life-year (QALY), were calculated, and incremental cost-effectiveness ratios less than $27,534.25/QALY (3 × the per capita GDP of China, 2018) were considered cost-effective. One-way sensitivity and probabilistic sensitivity analyses explored the robustness of the model. Our base case model found that the total cost was $53,082.68 in the GP group and $45,482.66 in the TPF group. The QALYs were 6.82 and 4.11, respectively. The incremental cost-effectiveness ratio favoured the GP regimen, at an incremental cost of $2,804.44 per QALY. The probabilistic sensitivity analysis found that treatment with the GP regimen was cost-effective 100% of the time at a willingness-to-pay threshold of $27,534.25‬/QALY. In this model, GP was estimated to be cost-effective compared with cisplatin, fluorouracil, and docetaxel for patients with locoregionally advanced nasopharyngeal carcinoma from the payer's perspectives in the China.