A least squares geofence method that minimizes trigger misfires caused by data variability and location blunders and minimizes delayed/missed entry triggers generated under urban or indoor conditions. The least squares geofence method uses a weighted least squares (LS) model to compute a location estimate for a target device. A LS location estimate is used to determine if a target device is located inside or outside a predefined geofence. The present invention additionally comprises a Kalman filter geofence method that further improves the accuracy of entry/exit geofence triggers. A Kalman filter geofence method uses a Kalman filter to filter location data retrieved for a target device. Filtered location data is used to determine if a target device is located inside or outside a predefined geofence. A Kalman filter geofence method estimates velocity and heading information for a target device to generate accurate entry/exit geofence triggers for devices in fast moving mode.