Mesh slicing is one of the most common operations in additive manufacturing (AM). However, the computing burden for such an application is usually very heavy, especially when dealing with large models. Nowadays the graphics processing units (GPU) have abundant resources and it is reasonable to utilize the computing power of GPU for mesh slicing. In the paper, we propose a parallel implementation of the slicing algorithm using GPU. We test the GPU-accelerated slicer on several models and obtain a speedup factor of about 30 when dealing with large models, compared with the CPU implementation. Results show the power of GPU on the mesh slicing problem. In the future, we will extend our work and standardize the slicing process.