In a previous article, we presented the model, the responses and the parameters of a solar refrigeration plant using a battery storage BSS. The measurements collected on the BSS system in laboratory and rural sites (stand alone PV plant) at different climatic conditions have shown some problems coming from the battery and its regulator, such as maintenance, replacement, break-down and the limited battery life period. In order to surmount these problems, we thought of substituting the electric storage by a thermal or a latent storage LSS (accumulation of ice in a separate section). To have a good prediction, optimal control and dimensions of the LSS system, we have developed a simulation program ＇＇SIREF＇＇ by using an acquisition/treatment environment. The validation of ＇＇SIREF＇＇ required the design and the development of a real prototype LSS plant. By installing a data acquisition chain, we have acceded to have the measured PV-LSS responses which are used to identify the system parameters by the recursive least square algorithm (RLS). The measured and the simulated results have shown the importance of the thermal losses and the choice of the insulation thickness. Hence, a study of the energetic optimization was elaborated in order to reduce the total life cost TLC for the LSS plant.