This paper describes a method to recognize the current location of a user using a few simple sensors. The proposed method is based on the integration of incremental motions (e.g., walking) over time to determine location transitions. To handle uncertainty, a fuzzy reasoning method is introduced. The paper also describes a prototype location recognition system that uses a set of inexpensive and wearable sensors, a bi-axial accelerometer and a digital compass. The system uses a reliable motion recognition method featuring only two directional acceleration measurements. The effectiveness of the proposed method is demonstrated by experiments at six locations in an indoor environment.