A UKF-based algorithm for asynchronous data fusion of wireless sensor is proposed. The RNAT mechanism is used to identify redundant nodes in the wireless sensor network, and a data redundancy tree is constructed to implement redundant data removal. According to the result of deduplication, in the environment where the radius of each sensor’s detection range is equal, two nodes that detect the target information are arbitrarily selected by the four-circle positioning method to calculate the intersection of the boundaries of two circular detection regions, ?nding and approximating the target according to the iterative method. The preconditions that the original sensors of di?erent sensors are independent of each other and the di?erent original quantity measurements of the same sensor are independent of each other are set, and the measured values of each channel are calculated. The unmeasured Kalman ?lter is used to update the measured value in the form of ?ltering, the Kalman ?lter gain matrix is introduced, and the asynchronous data positioning result is combined to realize data fusion. The experimental results show that the data utilization after fusion is higher than the current results, the algorithm short time-consuming and low energy consumption, and has high data fusion precision, the accuracy of the whole fusion is more than 90%.