Cold surge (CS) events are the most serious extreme cold events in winter in China, causing large economic losses and casualties. The occurrence of CS events has slightly increased since the 1990s. However, the possible future changes in these events remain unclear, and quantifying robust projected changes in CS events is important for developing adaptation and policy planning. Here, we project the occurrence of CS events and strong CS (SCS) events using the weighted multi-model ensemble (MME) of the Coupled Model Intercomparison Project 6 (CMIP6) through the application of the rank-based weighting (RBW) approach under three shared socioeconomic pathway (SSP126, SSP245, and SSP585) scenarios. The corresponding weights of each model were obtained depending on the comprehensive historical performance from three aspects: climatology, spatial variation, and interannual variability. The results show that the RBW approach can reduce the relative bias by approximately 50% compared to the unweighted MME. The occurrence of CS and SCS events shows a decreasing trend during 2015–2099 over northern China under the three SSP scenarios. There are also robust change projections during the long-term (2080–2099) and 2015–2099 periods under SSP245 and SSP585, especially in the NEC region, which exhibits a signal-to-noise ratio (SNR) that is >1. However, the occurrence of SCS events shows slight increases of 1.18% and 3.55% over northern China (notably western and eastern Northwest China) during the near-term (2020–2039) under SSP126 and SSP245, respectively. Obvious reductions in projected uncertainty are widespread throughout northern China after applying the RBW approach compared to the unweighted MME, which mainly depends on the scenario, region, and term variation. Then, a robust decreased frequency may contribute to projected changes in large-scale atmospheric circulation (a more positive AO and weaker SH and EAT) under SSP585. Our results emphasize that the weighted MME can be taken into account when projecting future extreme climate change in some areas to enhance reliability.