This research was carried out to estimate the CARDINAL temperatures of Ambrosia psilostachya seed germination and evaluation of some nonlinear regression models. The experiment was performed at Tehran University Weed laboratory in 2022. In this experiment, the seeds were exposed to 8 temperature treatments (5, 10, 15, 20, 25, 30, 35, and 40ºC) with 3 replications in a completely randomized design. The germination rate, seed germination percentage and mean germination time of seeds were measured. CARDINAL temperatures of seed germination were estimated using four regression models including segmented, dent-like and beta (four and five parameters) models. The performance of models and statistical indices was compared by RMSE (root mean squared of error), R2 (coefficient of determination) and AICc (corrected Akaike Information Criterion). Highest germination percentage was obtained at 15, 20 and 25°C (60.3%, 63.7% and 63%, respectively). The Highest germination rate (0.13d-1) and the lowest mean germination time (7.48d) were obtained at 25°C. Due to lower AICc and RMSE, the beta model (five parameters) was the best model for predicting the response of germination percentage to temperature in A psilostachya. Generally, base, optimal and ceiling temperatures were estimated 5, 25.5 and 42.13°C respectively by beta five-parameter model.