As a critical regulator of the global climate system, the Atlantic Meridional Overturning Circulation (AMOC) has attracted huge attention since its bimodal state change could abruptly alter climate. The subpolar sea surface temperature (SST) has been used as a fingerprint to predict the state change of AMOC. Although the subpolar SST agrees well with the variability of AMOC in recent years, it is under debate whether the subpolar SST can represent the abrupt state change of AMOC since it is not the domiant factor for deep convection. In contrast, it is well-known that the freshwater flux in the subpolar North Atlantic is a key to AMOC stability. To foresee the AMOC collapse based on salinity, we investigate the relationship of SST and SSS with various AMOC mean states from the Paleoclimate Modeling Intercomparison Project phase 4 (PMIP4). It is found that SST does not agree with changes in AMOC mean state under various climate conditions, while the sea surface salinity (SSS) within the subpolar North Atlantic (SPNA) does. Our results indicate that about 4.0 PSU drop in SPNA SSS is required for AMOC collapse to completely shut down. Based on reanalysis products and observational data, it is worth noting that no significant salinity trend was observed in the SPNA in recent years, implying the stability of deep water formation in recent decades. More attention should be paid to monitoring salinity variations in the SPNA and studying the anthropogenic-induced ice melting and alterations in the water cycle in high latitudes.