Many Arabic character recognition systems have been proposed since the early eighties. Most systems reported high recognition rates, however, they overlooked a very important factor in the process; the speed factor. In this paper, a high-performance Arabic character recognition system is introduced. The goal of the system is to maximize both accuracy and speed. The goal has been achieved through developing a high-accuracy sequential Arabic character recognition system and, then mapping it into a multiprocessing environment. Experimental results show that the multi-processing environment is very promising in enhancing a sequential Arabic character recognition system performance.