Image saturation is a challenging problem for three-dimensional shape measurement of objects with shiny surfaces using fringe projection profilometry technology. This paper presents a new method based on projection of multiple fringe patterns, providing an effective solution to the problem. First, a set of phase shifted sinusoidal fringe patterns with different intensity levels are projected to the object of interest, and the reflections are captured by the camera. With the captured images a set of masks are created, which are able to pick up the unsaturated data from the captured images. The unsaturated data are then combined based on maximal Signal-to-noise Ratio (SNR) scenario, and the combined images are used to retrieve the phase information. Comparing to existing methods, the proposed technique is advantageous by improved measurement accuracy due to the improved SNR. The performance of the proposed method is verified by the experiments.