An optimization and a statistical analysis are presented in order to obtain a mixture based on commercial and inexpensive components, which represents the enthalpy behavior of the package described in UNE-EN441 standard. The enthalpy of the M package, which represents the thermal characteristics of lean meat used in test of refrigerated display cabinets, is estimated as a function of temperature. The package is used to analyze the heat transfer process in packages undergoing cooling or freezing conditions. The package is composed of hard wood sawdust, brine and a preservative. The points required for experimentation are determine according to D-Optimal experiment design, taking into account the concentration of each component. The β’s parameters for the semi-theoretical non linear correlation of the modified package enthalpy are estimated with nonlinear least square approach. Gauss-Newton algorithm with Levenberg-Marquardt modifications for global convergence is used. Correlation for the enthalpy package is presented and suggestions for future experimentation are presented in order to increase the accuracy of the regression.