This paper presents a high-precision distance prediction system based on neural networks, integrating multiple ranging methods from Bluetooth 6.0 channel sounding to improve ranging accuracy. By fusing data from various ranging methods in different environments, we propose an attention-based multi-method fusion backpropagation neural network (AMF-BPNN) to improve the ranging accuracy. The proposed method is low complex and effectively extract the main features from different ranging data to attain the accurate range estimation. Experimental results demonstrate that the AMF-BPNN model achieves high-precision ranging performance with an average absolute error of 0.22 m within a range of 1–30 meters, showing significant improvement over single ranging methods.