Nowadays, DFP (Decimal Floating-point) is widely used in financial fields such as tax calculation, currency conversion and other areas where precise arithmetic is needed. Binary arithmetic, although widely used in current ALU (Arithmetic Logic Unit)s, has some limitations when performing correct decimal arithmetic. Consequently, DFU has drawn more and more attention in recent years. Due to the increasing demands for DFUs, IEEE 754-2008 formally defines three decimal DFU formats for both industry and research areas. More effort should be made on the spread of DFUs.In this thesis, a hardware based radix-100 divider is designed and implemented.Instead of using popular SRT (Sweeney, Robertson, and Tocher) division algorithm, selection by truncation algorithm is utilized. As a high-radix decimal divider, radix-100 divider can generate two quotient digits in each iteration. This is the major advantage of high-radix decimal divider compared to the decimal dividers. Besides, a compensation method is utilized to reduce the cycle time and the time consumed on the "multiples selection" module. Decimal carry-save adders and decimal carrypropagate adders are reused to reduce the overall area.The radix-100 divider is proven to be faster (3%) than the current decimal dividers, although the ratio is not outstanding. Meanwhile, the radix-100 divider consumes a larger area than the decimal dividers. It is expected to be a good start for the radix-100 divider. By applying more techniques in the future, the performance (latency) of the radix-100 divider is very likely to be much better than the decimal dividers.iii