Three different accelerated life tests (ALT) were designed by computer simulation to investigate their practical applicability to quantify kinetics of quality loss in frozen stored foods. Heat transfer and quality degradation inside a green bean were simulated, using a spectral finite element method (SFEM), to develop pseudo-experimental data. Temperature fluctuations inside a refrigerator were simulated, by a piecewise linear stochastic differential equation, and integrated into the SFEM program. Thereafter, the simulated data was treated by non-linear regression analysis to estimate the kinetic parameters. The different ALT tests were then compared in terms of precision and accuracy. This study shows that temperature fluctuations, inside a refrigerator, influence the accuracy of the kinetic estimates, and if the temperature spectrum is used to derive kinetic estimates, it is possible to apply accurately ALT methodologies to frozen foods. (C) 2003 Elsevier Ltd. All rights reserved.