Instrumental meteorological observations are essential for analysing past climate variability and reconstructing historical climate. However, many long instrumental climate records, some extending back to 1658 or 1659, contain inhomogeneities (artificial shifts) caused by changes in station location, instrumentation, observing practices, and environmental changes. These non-climatic influences reduce the consistency of long-term climate series and therefore require homogenization before the data can be reliably used for climate analyses.This article presents HOM-HCLIM, a homogenized version of the published global multivariable monthly instrumental climate database HCLIM (https://doi.org/10.1594/PANGAEA.940724). Three automated homogenization methods (CLIMATOL, BART, and PHA) were applied and evaluated using the HCLIM temperature records. Based on the evaluation of the three methods, BART was selected to produce the final dataset.HOM-HCLIM contains 2,892 homogenized monthly temperature series spanning the period 1757–2020. The dataset provides improved temporal consistency while preserving long-term climate variability and is intended to support studies of historical climate variability, climate reconstruction, and long-term climate change.