The 1990 Canadian long-range transport of air pollutants and acid deposition report divided North America into 40 sources of emission and 15 sensitive receptor sites. For the purpose of national policy making and international negotiation, the use of these large sources and few receptors may prove adequate. Due to inadequate information regarding cost of reducing emissions from each point source, it was felt necessary to design a method to generate cost functions for emission regions. The objective of this study was to develop aggregate cost functions that relate the cost of SO2 emission reductions to the amount of reduction achieved. The cost curves generated presume the application of control technologies to achieve a mandated regional emission reduction in the year 2000. The study has also assumed that trading will take place among plants within a region. The emissions inventories (GECOT and AIRS for the USA and RDIS for Canada) were used as the major source of data for the study. Cost functions were derived for forty emission regions. The functional forms that best fits estimated costs are either quadratic, power or linear in specifications. Furthermore, the cost functions indicted substantial variation (differences in marginal costs of removal) across emission regions. Preliminary analysis using Environment Canada’s Integrated Assessment Modelling platform indicated that strategies that make use of these functions and environmental goals will cost the industry and government the minimum amount compared to those that relay on quantitative emission reductions. Considering the findings of studies that indicated exposure of several watersheds to excess depositions of SO2 compared to critical loads, policy makers should examine ways of reducing emissions beyond what is already committed for the year 2005 or 2010. Future work will investigate interregional trading, especially between the bordering states of the USA and provinces of Canada based on these cost functions.