As an essential requirement for surveillance systems, target tracking has been studied extensively. Most of the tracking schemes are based on trilateration, which requires each point in the monitoring area to be covered by at least three anchors. However, due to the inadequate deployment of costly anchors and environment constraints, the target might not always be detected by three or more anchors simultaneously, resulting in intermittent localization failures and performance degradation. To address this issue, this paper proposes a tracking method called Bubble Trace (BT) for insufficient anchor coverage and asynchronous networks. By fully extracting the location information embedded in dual, single and zero anchor coverage, we develop a bi-directional bounding algorithm to offer the bubble-shaped regions that indicate the possible locations of the target. Moreover, instead of separately estimating each position point of the target, we construct the trace by finding a maximum-likelihood path in a graph. The design is evaluated through extensive simulation and a test-bed experiment with 20 MicaZ nodes. Results show that the proposed scheme improves the tracking accuracy without using additional hardware under insufficient anchor coverage.