This paper reports on the development and application of a method for the numerical prediction of temperatures within passive cooling components. In previous studies the basic idea was presented in an alternate form for the prediction of earth-contact heat transfer. The new method is demonstrated to be fast, accurate and flexible as a result of incorporating elements of the response factor method into a finite volume technique based numerical model. Initially, a 'pre-processing' procedure is required to generate a certain number of hours for use as a time series by the response factor technique in the second stage of the method. It is here shown that even a reduced number of temperature response factors, e.g. 50 h, is sufficient to obtain accurate predictions of the component's hourly temperature profile up to 1 year ahead. This study develops the equations addressing temperature profiles in structural components. The 'state of the art' of the presented method corresponds in the way that the conductive temperature response factors are calculated being numerically in the three-dimensional space. The method solves the three-dimensional earth-contact temperature profiles, which interact with indoors and outdoors temperature profiles. Once the numerical temperature response factors time series of an earth-contact component's grid node have been generated then its future thermal performance due to any surrounding temperature variation can be predicted fast and accurately. The structural earth-contact passive cooling components' response factors are generated by a three-dimensional numerical model, with no need of past experimental data and stored to be used at any time in future. The method is given the name NGRF (Numerical Generation of Response Factors). The way that the temperature response factors are determined in the three-dimensional space targets in the improvement of the prediction of earth-contact temperature profiles. (C) 2008 Elsevier Ltd. All rights reserved.