The theme of this Master Thesis is modeling and simulation of the musculoskeletal system. The purpose is to explore new ways to diagnose the human body, and do so effectively by automation. This Master Thesis investigates the problem of automatic segmentation to obtain geometry of bone, muscle, ligaments and tendons. In addition, it includes a research on how to generate simulation input in a Knowledge Based Engineering System (KBE). The scope has been to generate simulation input from a medical image. Simulation input includes bone, muscle, ligament and tendon geometry as well as mechanical properties for these tissues. Due to extent and complexity this Master Thesis has focused on extraction of bone data from the medical image and geometry generation as the simulation input. A segmentation program has been developed in Matlab to obtain bone as the only tissue from medical images. The medical images used in this Master Thesis have been taken with CT. Another program has been developed in a KBE framework, Technosoft's Adaptive Modeling Language. The segmentation program is able to segment bone tissue and present the result in 3D. However, it requires user input to control the segmentation process. The AML program generates geometry for tibia, femur, pelvis and the spine. It also generates cartilage to represent the intervertebral discs. The developed programs are platforms for further research and development, and has proven Matlab and AML as viable tools for continued work.