In order to improve the positioning accuracy of the fusion positioning algorithm of UWB and IMU, a particle filter positioning algorithm based on improved adaptive weighted data fusion was proposed. First, the minimum variance estimation theory in the adaptive optimal weighted fusion algorithm was used to adjust the particle distribution weight in particle filtering. The threshold was used to limit the observed variance to avoid divergence of the observed variance due to the actual environment. Using the observation noise covariance and measured values, the optimal weighting factors for each sensor were obtained in the RMSE limit interval after particle filtering to avoid the problem of algorithm divergence caused by weak or missing sensor signals. Finally, A comparative experiment of UWB and IMU fusion positioning was carried out. The experimental results show that the algorithm improves the positioning accuracy by more than 15% compared with the EKF fusion algorithm.