Image dehazing utilizes a very complex algorithm that requires intensive filtering and floating-point arithmetic operations. Consequently, processing speed is the most significant bottleneck in its application in some vision tasks such as mobile platforms. In this paper, we propose an optimized single image parallel processing dehazing algorithm for mobile platforms and implement it on a Windows Phone device based on GPU rendering technology.