Real-time systems surround every facet of our lives. They can be found in anything from everyday objects like mobile phones and washing machines to objects critical to life and infrastructure including heart rate monitors and nuclear power plants. As time progresses these systems are becoming ever more complex. To cope with the increase in complexity, developers and researchers are turning to model driven development as a solution. One modeling language aimed specifically at real-time systems is UML-RT. This thesis proposes an algorithm that for a significant subset of UML-RT is able to provide a worst-case execution time analysis for capsules at a model level. Having access to the worst-case execution times allows developers to at an early stage reason about a given system. This allows for better resource allocation as well as the ability to perform scheduling analysis. Development of the algorithm was performed iteratively using the constructive research approach. We began by first gaining an understanding of the theory. We then successively developed a theoretical algorithm selecting one or a few UML-RT entities at a time. With each iteration the algorithm was redefined to incorporate the new entities. At the end of the development, we created an implementation of the algorithm as an Eclipse Modeling Framework plug-in using Java. We then created a set of hard coded capsule models which were used to validate the algorithm.