Uncertainties are involved in every part of a solar thermal system due to tolerances in design parameters, tolerances in system parts, and the random nature of variables like solar radiation and thermal demand. Current design practices in solar engineering do not incorporate these uncertainties into the design process, despite the fact that their magnitude and combined effect are often large enough to make the result of the calculations very dependent on the assumptions about the true values of the input data. This paper proposes the application of reliability analysis methods to design solar thermal systems, in order to take into account the true stochastic nature of the problem. This approach allows designers to quantify the probability that the system will achieve the desired performance in the face of uncertainty. The methodology followed in the paper consists in applying uncertainty bands to the input data. Monte Carlo techniques are then used to propagate these uncertainties, calculate the reliability of the system and determine its sensitivity to each uncertain factor. The practical application of the proposed design methodology is illustrated with a case study, in which a typical solar hot water thermal system is dimensioned. (C) 2011 Elsevier B.V. All rights reserved.