Laser 3D curve positioning projection is a critical technology for aviation composite material layup processes. This study presents a novel 3D curve positioning and projection system based on a galvanometer scanning mechanism, consisting of a laser source, galvanometer, photosensitive sensor, focusing lens assembly, beam splitter, and auxiliary optics. During high-speed scanning, the photosensitive sensor detects reflected light intensity signals from the retroreflective target surface, while the galvanometer provides real-time control signal feedback. The 3D curve positioning and projection system acquires these two signal data types and utilizes a single hidden layer feedforward neural network (SLFN) to establish the mapping relationship between the input signals and the output laser lines, with calibration completed by solving parameters in the network model. By applying the non-perspective n-point (NPnP) algorithm, the system achieves target positioning and projects pre-designed patterns on the surface without requiring external measurement devices or precision optical component assembly. Experimental validation through retroreflective target positioning and aircraft composite panel contour projection confirms the system’s effectiveness.