The effect of temperature on the development of two Trichogramma species, Trichogramma japonicum Ashmead and Trichogramma ostriniae Pang et Chen, was studied in the laboratory under 7 constant temperature gradients (10, 15, 20, 25, 30, 35, 37℃). The results indicated that both the developmental period of T. japonicum and T. ostriniae decreased with the temperature increasing from 10 to 37℃. The developmental periods decreased rapidly from 10 to 30℃, which were 57.86, 28.75, 12.88, 9.50, 7.75d for T. japonicum and 61.13, 28.50, 13.00, 9.75, 8.05d for T. ostriniae respectively under 5 constant temperature gradients (10, 15, 20, 25, 30℃). The differences of the developmental period were significant among these 5 temperature gradients. But the developmental periods decreased slowly with the temperature increasing from 30 to 37℃, which were 7.75, 7.45, 7.95d for T. japonicum and 8.05, 7.50, 7.90d for T. ostriniae respectively under 3 constant temperature gradients (30, 35, 37℃). There was no significant difference among these 3 temperature gradients. The effects of temperature on the development of the 2 Trichogramma species were consistent with the linear regression and Wang-Lan_Ding models. The developmental thresholds of T. japonicum and T. ostriniae developing from egg to adult were 5.23℃ and 5.24℃ respectively, and their generational effective accumulated temperatures were 212.77 and 217.39 degree\\days respectively. Both the developmental threshold and the effective accumulated temperatures were calculated from the linear regression model. The lowest, optimum and highest lethal temperatures estimated by Wang-Lan-Ding model were 7.96℃, 18.91℃, 38.99℃ respectively for the development of T. japonicum, and 6.97, 18.87, 38.99℃ respectively for T. ostriniae. The developmental rates of both species increased with the temperature increasing from 10 to 37 ℃, rapidly from 10 to 30℃, but slowly from 30 to 35℃. When temperature increased beyond the highest lethal temperature, the developmental rates did not increase anymore at the end. The design of the consistent temperature gradient and the mathematical model selection in the “temperature-development” study of Trichogramma were also analyze. The developmental threshold obtained from the linear regression model may provide introduction for the parasitoids reservation in lower temperature. Integration of the linear regression and Wang-Lan-Ding models is sufficient to get the result from the “temperature-development” study, because the generational effective accumulated temperature can be provided by the linear regression model, while Wang-Lan-Ding model can reflect the change of the developmental rate under the lowest to the highest lethal temperature. The number of the constant temperature gradients must not be less than the number of the parameters of Wang-Lan-Ding model, and the lowest temperature of the constant temperature gradients must be close to the lowest lethal temperature, while the highest temperature is close to the highest lethal temperature when using this model to theoretically simulate the relationship between temperature and development.