In computer arithmetic research field, one of the challenging problems is how to improve the performance of fundamental arithmetic operations. Most researchers have proposed various number systems which are suitable for a certain type of computation. However, many number systems have the same limitation in performance when operating with the multiply and divide calculation. Therefore, the number systems that perform well in those operations have been proposed. Logarithm number system is one of the number systems which have an advantage in multiplication and division. Unfortunately, this number system has its limitation in addition and subtraction because it requires a look-up table. Hence, many researchers focus on how to reduce the size of a look-up table. This thesis proposes an improvement version of the double dimension logarithmic number system called an extended double dimensional logarithmic number system. By our proposed the addition and the subtraction algorithm, this number system shows a significantly reduction in the usage of look-up table comparing with the classic double dimensional logarithmic number system. Fundamental arithmetic operations such as multiplication and division are also introduced in this work. We also propose a novel approach to solve the accuracy problem in this number system.