A diagnosis procedure is an algorithm to detect and locate (isolate) faulty components in a dynamic process. To achieve diagnosis, redundancy has to be included in the system. This redundancy can be either hardware redundancy or analytical redundancy. Methods based on analytical redundancy need no extra hardware. The redundancy is generated from a process model instead. In this thesis, methods based on analytical redundancy are used. A mathematical model and a controller are derived for an inverted pendelum. Two diagnosis methods, parity equations and parameter estimation, are implemented on the pendelum. Different faults are then applied to the pendelum. Weights on the rod and on the cart of the pendelum corresponds to component faults, constants added to the output and input signals corresponds to sensor faults and actuator faults respectively. The diagnosis methods are studied in order to see how good the methods are compared with each other and to results in scientic articles using similar methods.