A wearable mobility monitoring system (WMMS) could be a valuable device for rehabilitation decision-making. A proof-of-concept system is developed that uses the BlackBerry 9550 as a self-contained WMMS platform. An integrated tri-axial accelerometer, GPS, and timing data are processed to identify the mobility change-of-state (CoS) between standing, walking, sitting, lying, stair climbing, going up or down a ramp, riding an elevator, and riding in a car. Following feature extraction from the sensor data, a decision tree is used to distinguish the CoS. In the complete system, real-time CoS identification on the smartphone will trigger video capture for improved mobility context analysis. Preliminary evaluation involved collecting three trials from one subject while he completed a continuous circuit that incorporated all target mobility tasks. The average sensitivity is 89.7 % and the specificity is 99.5 % for walking-related activities. The sensitivity is 72.2 % for stair navigation and 33.3% for ramp recognition, since accelerations for a ramp gait are similar to those for a walking gait. These results provide insight into algorithms and features that can be used to recognize CoS in real-time.