A multiple-precision arithmetic system consist of integer, rational number floating point number and complex number has been developed using C++ language. These multiple-precision numbers are represented as classes in C++ language. Therefore it is easy to convert C or C++ source codes to multi-presision ones because these numbers can be treated as the predefined numbers such as int, float and double. Simple examples is given to shown how to use this system. We also discuss the problems to implement multiple-precision arithmetic system in C++ language.