This paper proposes a technique combining an adaptive fringe projection algorithm and a curve fitting algorithm for the three-dimensional (3D) reconstruction of highly reflective objects. First, an adaptive projection algorithm is proposed to avoid image saturation. The phase information obtained by projecting vertical and horizontal fringes is used to establish the pixel coordinate mapping between the camera image and the projected image, and the highlight area in the camera image is found and linearly interpolated. By modifying the mapping coordinates of the highlight area, the optimal light intensity coefficient template of the projection image is calculated, the coefficient template is mapped to the projector image, and it is multiplied by the standard projection fringes to obtain the adaptive projection fringes we need. Secondly, after obtaining the absolute phase map, the phase at the hole is calculated by fitting the correct phase values at both ends of the data hole, and the phase value closest to the actual surface of the object is obtained by performing a fitting in the horizontal and vertical directions. Multiple experimental results show that the algorithm can reconstruct high-quality 3D shapes for highly reflective objects, with high adaptability and reliability in high dynamic range measurements.