Aiming at the problems of susceptibility to noise and low stability of fourrotor UAV with single attitude sensor, the gyroscope array is used to form multinode, antijamming and stable multiattitude system. A flight attitude measurement system of fourrotor UAV based on the new gyroscope array is proposed. The measurement array is composed of multiple micromechanical electronic systems with low accuracy, which improves the accuracy and stability of the system data. At the same time, a corresponding BP network data fusion algorithm based on neighborhood search is proposed, which solves the problem that the traditional BP neural network requires accurately giving the output value. The BP neural network model was used in the data fusion processing of the gyroscopearray. The experiment results show that the multigyroscope array system designed in this paper obviously improves the antinoise performance compared with the singlegyroscope system. Compared with the traditional linear weighted fusion algorithm, the proposed algorithm increases the support degree by 92% and reduces the residual error by 442%. The practicality experiment shows that the proposed method has practical significance in improving the flight stability of fourrotor UAV.