The ability of milk and concentrated milk to withstand a defined heat treatment without noticeable changes such as flocculation of protein is commonly denoted as heat stability. A heat treatment that exceeds the heat stability limit of milk or concentrated milk, which has a much lower heat stability, may result in undesired changes, such as separation of milk fat, grittiness, sediment formation, and phase separation. Most laboratory-scale batch heating methods were developed in the early 20th century to simulate commercial sterilization, and these methods have since been standardized. Heat stability studies have been motivated by different objectives during that time, addressing different processing issues and targets in the framework of available technology, legislation, and consumer demand. Although milk hygiene has improved during the last couple of decades, rendering milk less sensitive to coagulation, different standard methods suffered from poor comparability of results, even when comparing results for the same milk sample, indicating that unknown procedural steps affect heat stability. The prediction of heat stability of concentrated milk from the heat stability results of the corresponding unconcentrated milk for rapid quality testing purposes has been difficult, mainly due to different experimental conditions. The objective of this study is to review literature on heat stability, starting from studies in the early 20th century, to summarize the vast number of studies on compositional aspects of milk affecting heat stability, and to lead the way to the most recent work related to compositional changes in concentrates produced by membrane concentration and fractionation, respectively. Particular attention is paid to early and most recent developments and findings, such as the application of kinetic models to predict and limit protein aggregation to assess and describe heat stability as a temperature-time-total milk solids continuum.