Abstract: In this study, a rat morphine drug discrimination model with a fixed ratio (FR) of 10 (FR10) was established using different methods to explore which methods can shorten the modeling time and test the dose-response relationship and median effective dose (ED₅₀) value. Animal welfare and experimental procedures are in accordance with the provision of the Animal Ethics Committee of Shanghai InnoStar Bio-tech Co., Ltd. Forty rats were initially shaped to press lever under a fixed-ratio schedule of food reinforcement. The animals that were successfully trained under a FR10 schedule of food reinforcement were divided into two groups, namely the single-lever + double-lever training group 1 and the double-lever training group 2. In each group, rats were trained to discriminate morphine at 5.6 mg·kg^-1 from saline by the intraperitoneal route. After training, different doses of morphine were used to substitute for training dose of morphine, the dose-response curve for morphine were identified in rats, and the ED₅₀ value was calculated. The results showed that, in food training phase:34 rats successfully entered the discrimination training during food training; in discrimination training phase:14 animals in group 1 met the discrimination training standard for the first time, which took about (40.71±2.93) days, and there were 13 animals in group 2 that met the discrimination training criteria for the first time, and it took about (51.15±2.55) days. It can be seen that the method of single-lever + double-lever training is better than single-lever training, and the difference is significant compared with group 1 (P<0.05); in generalization test phase:there are 17 rats completed morphine generalization test, and the percentages of morphine-lever responses produced by the generalization test of different doses of morphine (0, 0.1, 0.5, 1, 3, 5.6, and 10 mg·kg^-1) were (9.56±3.13)%, (9.01±5.83)%, (13.82±7.95)%, (29.04±10.13)%, (41.70±10.65)%, (85.36±7.16)%, (94.56±2.76)%, respectively. The results showed that the discriminative stimulative effect induced by morphine dose between 0-10 mg·kg^-1 increased in a dose-dependent manner, producing a good dose-response curve, and the ED₅₀ value of morphine was 4.74 mg·kg^-1 by linear fitting. The above results showed that, the FR10 morphine drug discrimination model has been successfully established using different methods; the single-lever + double-lever training method is better than the single-lever training, and can relatively shorten the discrimination training cycle.