Purpose: This study examines the net social benefit of a Supported Employment Program for People with Disabilities (2003-2004) in a Social Welfare Foundation. It is hoped that the findings of this study will serve as a reference to the government and private social welfare agencies implementing such programs. Methods: This study analyzed a Supported Employment Program for People with Disabilities (2003-2004) in a Social Welfare Foundation using a longitudinal cost-benefit analysis method (focusing on monetary costs and benefits). Three groups were analyzed: participants in the program, others (not participating in the program), and society (participants and others). Two costs were studied: administrative cost (including the payment for staff and facility), and opportunity cost (i.e. the assistance/allowances participants got before joining this program). In addition, three items of monetary benefits were studied: working salary, payroll tax (which is negative for participants but positive for others), and fringe benefits (i.e. labor insurance and national health insurance). Results: The administrative costs for the program for the years 2003 and 2004 were the same: NT$1,321,272 per year (a total of NT$2,642,544 over the two years). There were 70 participants in 2003 and 2004, and 51 of whom had at least at one time been employed. Therefore, the total administrative cost for others and society was the same as NT$18,875.31 per participant per year, and zero for participants. The opportunity costs were NT$0~NT$78,000 per participant per year, -NT$78,000~NT$0 for others, and zero for the whole society. Thus, the total social cost of this Supported Employment Program for People with Disabilities (2003-2004) in a Social Welfare Foundation was NT$18,875.31 per participant per year. On the benefit side, the total salary of the 51 working persons was NT$9,897,268; NT$3,272,748 and NT$6,624,520 for 2003 and 2004 respectively. Therefore, the salaries were NT$70,694.77 per participant per year for participants, zero for others, and NT$70,694.77 for the whole society. The estimated payroll taxes were NT$4,241.69 per participant per year for participants, NT$4,241.69 for others, and zero for the whole society. Fringe benefits are shown as positive/zero, for participants and negative/zero for others. The social fringe benefit was zero. That is to say, the total social monetary benefit was NT$70,694.77 per participant per year. Therefore, the net social monetary benefit just after this program was NT$51,819.46 per participant per year. The benefit cost ratio was 3.75. At this time, the net monetary benefits for the participants and others (tax-payers) were more likely to be positive. Considering sensitivity analysis, the estimated social monetary benefit stream in the future is NT$8,184,857 (in 2004 dollars). Divided by 70 participants, the estimated social monetary benefit stream in the future is NT$116,926.53 (in 2004 dollars) per participant. That is to say, it is highly possible that the net social monetary benefit will be NT168,745.99 per participant from now to the future. The benefit cost ratio will be 9.94 then. Furthermore, the net monetary benefit for participants will be definitely positive. Conclusions: This study is the first to use longitudinal cost-benefit analysis method to analyze a supported employment program for people with disabilities in Taiwan. The findings suggest that supported employment programs for people with disabilities can be efficient and equal for the whole society. The results suggest that firstly, the government should keep implementing supported employment programs for people with disabilities, and make a positive correlation between the evaluation result and delegating numbers and fees for an agency. Secondly, government and private sectors should cooperate to provide adequate employment opportunities and comprehensive working ability evaluations for people with disabilities based on their disability categories and levels. Thirdly, central government should regulate the unified standard of closure due to stable employment in the whole country, and ask the delegated agencies to follow the client situation after his/her closure at least half a year. It is all suggested that a longer longitudinal study can be carried out to try to quantify and even monetize the non-monetary benefits in the near future.