The purpose of this study is to integrate experimental design, system simulation, and GIS to develop an algorithm which can help grouping signals for better traffic operations. By definition, signal grouping is to cluster neighboring signals into a coordinated plan to smoothen traffic operations on roads. Many thresholds have been employed to determine if signals can be grouped together. Those thresholds include block length, road capacity, turning movement, number of lanes, traffic flow direction, street corss section layout, etc. The application of those thresholds in signal grouping in the past tended to use the superimposing or the annexing method. Therefore, this study employs the approach of orthogonal array to test the threshold combination for signal groupin. By doing this, a test network is designed which can represent the characteristics of most commonly seen street networks. Different combinations of traffic flow, block length, etc. are then put together on the test network to generate the efficiency of the threshold combinations. To test the efficiency of a certain threshold combination, the CORSIM simulation program is used to produce the MOE (measure of efficiency), and the program of TRANSYT-7F is used to provide the timing plan for the signal network, It is therefore this research a well designed and thorough experiment test. During the testing process, the GIS plays the role to provide the network characteristics which enhances the interface of data provision and the display of output. In addition, the development of the Tainan City's signal GIS will later help to test the validation of the research result. Through coordination with ciyt signal control division, the result of the study is regarded as valid and solid by engineering judgement.