Abstract: By a theoretical investigation of simulated nonisothermal data, the inability to determine the rate order for degradation of drugs with ordinary nonisothermal experiment was discussed. Results indicate that both rate order n and activation energy E can change the curvature of the c-t curve and the change caused by higher n can be compensated by higher E in ordinary nonisothermal experiments. That means a suitable E can always be assessed to fit each of the definite n (0, 1, or 2 in our study) very well. As a result, a same set of c-t data can be well fitted by different combinations of n and E. This is why the rate order cannot be assessed in ordinary nonisothermal experiments. To overcome this deficiency, it is important that there have both temperature rising and lowering parts in one nonisothermal experiment. A new heating model, temperature rising and lowering program, was presented. With this new heating model, the kinetic parameters, including the rate order, can be obtained in one nonisothermal stability experiment. The ability of the new heating model to determine rate order is almost the same as that of isothermal experiments.