<p>Nonlinear inverse heat conduction problems are classical problems in combustion, thermal management and industrial processes. Different approaches have been developed for solving such problems. However, most approaches for resolving inverse heat conduction problems are highly dependent on system parameters, which introduces a significant amount of predictive uncertainty. However, temperature-dependent thermophysical properties are often difficult to be measured accurately, which also adds some challenges. To simplify the data processing procedure and enhance the accuracy of prediction, this paper attempted developing a scheme to predict surface heat flux under nonlinear heat conduction with avoiding the measurement of material’s thermophysical properties, and reduce the uncertainty contained in prediction results. Numerical simulations verified this method is not sensitive to the noise of in-depth temperature measurement. This approach will provide a novel strategy for solving nonlinear inverse heat conduction problems with high accuracy and robustness.</p>