This study proposes a solution technique for solving the linear goal programming (LGP) problems, which can be effectively run on microcomputers with at least 64K bytes of memory. Because of the limited memory of microcomputer, the technique employs the sparse matrix technique, based on the modified revised simplex (MRS) algorithm using the modified product form of inverse (MPFI). A new scheme which stores the basis and implements the MPFI technique, is developed. And the new reinversion scheme for basis is performed to reduce roundoff errors with less storage. The computational experiments with several test problems show that the designed technique requires reasonably small memory space with the acceptable reoundoff errors.