The Extended Reconstructed Sea Surface Temperature Version 4 (ERSST.v4) is an updated version of the monthly sea surface temperature (SST) dataset, providing improved accuracy and completeness. Key upgrades include updated input data from the International Comprehensive Ocean–Atmosphere Data Set (ICOADS) release 2.5, revised empirical orthogonal teleconnections (EOTs), updated SST quality control procedures, and improved SST anomaly (SSTA) evaluation methods. The dataset also incorporates updated bias adjustments for ship and buoy SSTs using the Hadley Centre Nighttime Marine Air Temperature dataset version 2 (HadNMAT2). These revisions significantly impact SST values, making them cooler north of 30°S and warmer south of 30°S compared to ERSST.v3b before 1940. Comparisons with other SST datasets like HadSST3 show that ERSST.v4 is cooler in the tropics but warmer in midlatitude oceans. The revised ERSST.v4 better represents El Niño/La Niña behavior in sparse observational periods before 1940 and is closer to satellite-based observations. The methodology involves using in situ data, NMAT for bias adjustment, and EOTs to reconstruct SST anomalies. The dataset includes improvements in data coverage, quality control, and bias adjustments, leading to more accurate SST reconstructions. The impact of these upgrades is discussed in detail, highlighting the importance of bias adjustments and the role of EOTs in capturing long-term SST trends and decadal variations. The study also addresses the challenges of data sparsity and the need for robust methods to handle uncertainties in SST reconstructions.The Extended Reconstructed Sea Surface Temperature Version 4 (ERSST.v4) is an updated version of the monthly sea surface temperature (SST) dataset, providing improved accuracy and completeness. Key upgrades include updated input data from the International Comprehensive Ocean–Atmosphere Data Set (ICOADS) release 2.5, revised empirical orthogonal teleconnections (EOTs), updated SST quality control procedures, and improved SST anomaly (SSTA) evaluation methods. The dataset also incorporates updated bias adjustments for ship and buoy SSTs using the Hadley Centre Nighttime Marine Air Temperature dataset version 2 (HadNMAT2). These revisions significantly impact SST values, making them cooler north of 30°S and warmer south of 30°S compared to ERSST.v3b before 1940. Comparisons with other SST datasets like HadSST3 show that ERSST.v4 is cooler in the tropics but warmer in midlatitude oceans. The revised ERSST.v4 better represents El Niño/La Niña behavior in sparse observational periods before 1940 and is closer to satellite-based observations. The methodology involves using in situ data, NMAT for bias adjustment, and EOTs to reconstruct SST anomalies. The dataset includes improvements in data coverage, quality control, and bias adjustments, leading to more accurate SST reconstructions. The impact of these upgrades is discussed in detail, highlighting the importance of bias adjustments and the role of EOTs in capturing long-term SST trends and decadal variations. The study also addresses the challenges of data sparsity and the need for robust methods to handle uncertainties in SST reconstructions.