Predicting critical heat flux (CHF) on engineered surfaces is essential for optimizing surface design in terms of thermal limit of energy conversion systems. Based on the mechanism of receding capillary flow, which mimics in-situ hydrodynamic behaviour of triple contact lines, a CHF model was developed as a function of the arithmetic roughness height, nanoscale surface area ratio, and apparent contact angle, all of which successfully separated the surface effects on the CHF. Without any adjusting constants, the present model mechanistically predicted 60 CHF data included in seven experimental groups (five from literature) within an error of +/- 20%, showing better accuracy than those predicted by existing wettability- and roughness-based models. The present model will be useful in optimizing micro/nano scale design of surface structure for improved thermal safety of generic thermal applications demanding high heat flux boiling. (C) 2019 Elsevier Ltd. All rights reserved.