A new numerical technique named stochastic interval finite volume method is proposed for the steadystate temperature field prediction with hybrid uncertainties in material properties, external loads and boundary conditions. Random variables are used to treat the uncertain parameters with sufficient information; whereas, interval variables are used to quantitatively describe the uncertain parameters with limited information. Based on the perturbation theory and random interval moment method, the expectation and variance of the interval bounds of the temperature field are calculated effectively. By comparing the results with traditional Monte Carlo simulation, three numerical examples are given to verify the feasibility and effectiveness of the proposed method for solving steady-state heat conduction problems with hybrid uncertain parameters, pure random parameters and pure interval parameters, respectively. Crown Copyright (C) 2014 Published by Elsevier Ltd. All rights reserved.