The metered taxi fare calculation in Taiwan is based on an initial fixe d fare for the flag drop plus a variable fare related to the distance traveled and the time elapsed through meter incrementing until the end of the ride. Though this metered taxi fare structure has been set up in Taiwan, it is only executed in few cities in practice. The trades in most cities are still through negotiation between the passenger and the driver which unfortunately often results in pricing problems and quarrels. One of the major reasons is that the current taxi fare is difficult in reflecting the reasonable profit of the drivers in the actual operations.This research aims to develop a systematic approach in a time-space scheme to improve the current design of the taxi fare. We first analyze the relationship between metered taxi fare and trip distance from aspects of time and space.Then we use nonlinear optimization to develop formula for calculating metered taxi fare. Finally, we apply Monte Carlo Simulation to determine the best fare parameters. To evaluate the new approach in real world, we performed a case study using C computer language on a personal computer. The results show that the framework could be useful in actual operations.