This paper presents the application of multi-inputs single-output (MISO) models to estimate thethermophysicals parameters of a building. ARARMAX, Box-Jenkins and the general MISO models areused to identify the U value, the time constant and the equivalent solar surface of the building.Optimization-based prediction error method (PEM) algorithm is used to estimate model parameters.This approach has been tested to analyze a passive solar house in Tunis. The identified parameterswere compared to theoretical values; good results have been obtained for the tested building.