With the aim to validate and accurately evaluate the applicability of time-temperature indicator (TTI) application at variable temperatures, a prediction method based on the fuzzy reasoning was built. The method was on the basis of quality experience equation of the monitored products at constant temperature experiment, which could be chosen from polynomial equation, the n-th reaction kinetic equations or other equations according to the fitting coefficients. The key of this method was to build exact membership functions between arbitrate effective temperature and the constant temperature in order to obtain the predicted value at arbitrate effective temperature. The method was analyzed about Muscat Hamburg grape, Vitsab M25-2 and OnVu TTI through two fluctuant temperature experiments simulating temperature characteristics of table grape cold chain logistics. Triangle membership function was chosen in the prediction based on the fuzzy reasoning. The table grape quality predicted values based on the fuzzy reasoning and the kinetics model were compared with the actual measured values. Results showed that the TTI prediction method based on fuzzy reasoning at low fluctuant temperature made improvements (6.03 percentage points and 2.70 percentage points) vs TTI prediction method based on the reaction kinetics equations, whereas made no improvements at high fluctuant temperature. Therefore, TTI prediction method based on fuzzy reasoning could be chosen at low temperature based on the principle of merit.