Based on monthly data from 30 models in the Coupled Model Intercomparison Project Phase 6 (CMIP6), the leading mode of the inter‐model uncertainties in the projections of the summertime Northwest Pacific subtropical high (NWPSH) under the high emissions scenario (SSP5‐8.5) is identified by inter‐model empirical orthogonal function (EOF) analysis of the sea level pressure (SLP) over the East Asia‐Northwest Pacific region. The leading mode is characterized by positive SLP anomalies over the East Asia‐Northwest Pacific region. Inter‐model regression analyses reveal the relationship between the leading mode and historical sea surface temperature (SST). The inter‐model uncertainties of the NWPSH projections are closely related to the historical tropical North Atlantic (TNA) cold SST anomalies. The TNA cold SST anomalies result in increased precipitation in the subtropical North Pacific through westward propagated Rossby wave responses, further enhancing the subtropical North Pacific cold SST anomalies in the future through the reinforced convection‐shortwave‐SST negative feedback. Suppressed convection over the subtropical North Pacific strengthens the NWPSH through the anticyclonic Rossby wave response. The emergent constraints based on the observed TNA SST pattern reduce the uncertainties of the NWPSH projections by 32.5%, indicating a weaker NWPSH compared to the multi‐model mean. Cold biases of the simulated present‐day TNA SST in most CMIP6 models indicate that the multi‐model ensemble mean (MME) may overestimate the intensity of NWPSH in the warmer future. This study helps improve the accuracy of NWPSH projections, contributing to climate change adaptations for many important sectors impacted by NWPSH variation, such as agriculture, infrastructure, and energy.