Decision support systems (DSSs) in medicine are designed to aid medical professionals on making clinical decisions about prevention, diagnosis and corresponding treatment. When DSSs are applied to medical procedures, two sorts of predictions are possible: procedural (i.e. indications on what to do), and temporal (i.e. indications on what are the time restrictions). Clinical Practice Guidelines (CPGs) are statements that assist physicians making appropriate medical decisions during patient encounters. They are a set of assertions used to manage patients with a particular disease to improve quality of care, decrease unjustified practice variations and save costs. Clinical algorithms (CAs) obtained from CPGs are introduced to make the procedural knowledge explicit and formal. It is important to enable the latest clinical knowledge to be accessible and usable at the point of care, and therefore make significant contributions to safety and quality in medicine. Medical knowledge is used to assist patients suffering from one or several diseases. CAs could be explicitly given, or obtained with a knowledge management mechanisms. Among these mechanisms, there are some that aim at generating CAs from existing patients’ data for a particular disease. However, either explicitly given or generated CAs are atemporal, which means that there is no an explicit time labelling of the elements in the CA. Time plays a major role in medicine and therefore also in medical information systems. It is an important concept of the real world, which needs to be managed in different ways (events occur at some time points, facts hold during time periods, temporal relationships exist between facts and events) (Combi et al., 2010). If we want to overcome the gap of atemporal CAs it is necessary to define a time dimension and make also temporal knowledge (the indications on what are the time restrictions) explicit and formal. It has been proved that obtaining explicit temporal knowledge from physicians is often a difficult and time-consuming task regardless of the knowledge engineeringmechanisms or tools employed to simplify the process. As data saved in hospital databases are primarily time dependent, they can be used to obtain temporal constraints to define the time dimension of CAs. We have propose generation of temporal constraints considering patients’ data of a particular disease for atemporal CAs. We have defined two types of temporal constraints: macro-temporality and micro-temporality. Macro-temporality is defined as a constraint [tmin, tmax] on the time required to cross a particular edge of a CA, where tmin and tmax are the lower and the upper Temporal Knowledge Generation for Medical Procedures