High-precision and high-efficiency on-machine measurement is the premise for evaluating machining accuracy and ensuring machining quality of aerospace large thin-wall parts with large size, thin wall and weak rigidity. Multi-sensor data fusion is an important means to achieve high-precision and high-efficiency measurement of large thin-walled parts, however, the existing multi-sensor data fusion measurement methods rely on the explicit function reconstruction of curved surface, which is susceptible to the uncertainty of the measurement data and make it difficult to ensure the stability of the fusion results. A weighted residual fuzzy learning (WRFL)-based multi-sensor fusion measurement method for large thin-walled parts is proposed in this paper, in which, the residuals between different sensor measurement data are characterized by partition to obtain fuzzy-weighted fusion. Firstly, the low-precision point cloud is clustered and partitioned based on the high-precision data by probe measurement. Then the discrete residuals of lowprecision point cloud data in each partition are solved, and the residual sets are obtained by weighting the residuals in the cluster boundary region. The fuzzy set is finally established based on the discrete residuals to construct the high-precision fusion point cloud and realize the surface reconstruction. The experimental results demonstrate that the proposed method can significantly improve the surface measurement accuracy compared with the existing fusion measurement, and provides technical support for high-precision and high-efficiency measurement of large thin-walled parts.