Currently, the urgent task is to assess the small-sized maneuvering UAVs movement parameters. The location of an unknown UAV as a radio source can be determined using AoA measurements of the wireless sensor network. To describe the movement of a maneuvering UAV, a model is used in the form of a dynamic system with switching in discrete time. The values of switching variable determine type of UAV movement. To synthesize trajectory filtering algorithms, the Markov property of the extended process is used, which includes a vector of UAV movement parameters and a switching variable. The optimal trajectory filtering algorithm describes a recurrent procedure for calculating the a posteriori probability density function of an extended process. The optimal filtering device is multi-channel with feedback between the channels. To synthesize a quasi-optimal algorithm, linearized equations of UAV coordinates measurement in a Cartesian coordinate system based on AoA-measurements of a sensor network were obtained and an measurement errors analysis was performed. The quasi-optimal algorithm is obtained using the Gaussian approximation method of conditional a posteriori probability density functions and implements sequential processing of incoming measurements. It provides a joint solution to the problems of estimating UAV coordinates and recognizing of its movement type. Analysis of developed algorithm efficiency was carried out by Monte Carlo method. Shows the dependences of movement types recognition probabilities. A comparative analysis is performed with the Kalman filtering algorithm.