In this paper, we present a stochastic two-level optimization model whose upper-level problem depicts a wastewater treatment plant deciding on the size of compressed natural gas (CNG) filling stations and their locations. These upper-level decisions are integrated with operational decisions for the plant as well as downstream markets including agriculture, CNG transportation, residential natural gas, and electricity markets at the lower level. The two-level problem, expressed as a stochastic mathematical program with equilibrium constraints (SMPEC), is reformulated as mixed-integer linear program (MILP) using SOS1 transformations and linearizations. As a case study, the SMPEC is used to evaluate the options for CNG investment for a wastewater treatment plant located in the Washington, DC metro area. Our results indicate that the CNG produced from the wastewater treatment plant could meet approximately 20% of the expected total transportation demand in Washington, DC. In addition, CNG produced from the wastewater treatment plant could reduce CO\(_2\) emissions by a significant amount. The CNG benefits are traded off with less on-site wastewater-derived power production.