The net surface energy flux (Fs) is critical to the Earth’s energy budget and surface processes, but its simulation remains uncertain in global and regional scales. This study investigates simulated Fs biases and sources globally and in the Asian monsoon region (AMR) using CMIP6 HighResMIP atmospheric models. Globally, the multi-model mean can reproduce the observed global multiannual mean Fs. The majority of models overestimate the annual mean Fs, net surface shortwave radiation (SWs) and longwave radiation (LWs) but underestimate the turbulent heat flux (THF). In AMR, the Fs is predominantly upward during winter and downward during summer owing to the seasonal variation in SWs and THF. 95% of the winter Fs bias over AMR comes from THF primarily due to the latent heat flux bias. SWs and THF contribute 40%~90% and 70%~90% to summer Fs bias, respectively. The systematic biases of SWs and LWs can be attributed to biases in circulation patterns and cloud cover, while biases in THF are primarily influenced by the near-surface processes. The high-resolution models perform well in Fs, THF, and low-level circulation, particularly in DJF. The winter multi-model mean error is reduced by 21.5%~63.6% in Fs and 25.5%~76.7% in THF across three subregions of AMR. Seven out of nine high-resolution models show higher skill scores of winter Fs and THF than their low-resolution counterparts in SA, with corresponding model number being 8 (Fs) and 7 (THF) in both EA and WNP. This study reveals the advantages of increased horizonal resolution in the Fs simulation.