In this study, we modify the slack-based measure network data envelopment analysis (SBM network DEA) approach proposed by Kao (2014b) to calculate the efficiency of regional economic systems in Taiwan’s counties/cities during the period 2003~2014. For applying the SBM network DEA model to analyze the efficiency of the economic system, we divide the regional economic system of Taiwan’s counties/cities into a series of five stages— entrepreneurship, sustainability, operation, distribution, and consumption—and include the shared inputs and undesirable variables in the analysis framework. Our findings indicate that the average efficiency scores at the operation and consumption stages in the economic systems are the highest and lowest, respectively. We also observe a gap between the average efficiency at the entrepreneurship and operation stages between the urban and rural areas. Besides, gaps exist between the service and non-service counties/cities at the sustainability, operation, allocation, and consumption stages in Taiwan’s counties/cities. Finally, we identify wider efficiency differences among the regions in all the stages. Local officials responsible for the economic development can find out their own advantage and disadvantage stages and select the prior stages to improve the system performance according to our empirical results.