Numerical reservoir simulation has been used to describe mechanism of methane gas desorption process, diffusion process, and fluid flow in a coalbed methane reservoir. The reservoir simulation model reflects the response of a reservoir system and the relationship among coalbed methane reservoir properties, operation procedures, and gas production. This work presents a procedure to select the optimum well spacing scenario by using a reservoir simulation. This work uses a two-phase compositional simulator with a dual porosity model to investigate well-spacing effects on coalbed methane production performance and methane recovery. Because of reservoir parameters uncertainty, a sensitivity and parametric study are required to investigate the effects of parameter variability on coalbed methane reservoir production performance and methane recovery. This thesis includes a reservoir parameter screening procedures based on a sensitivity and parametric study. Considering the tremendous amounts of simulation runs required, this work uses a regression analysis to replace the numerical simulation model for each wellspacing scenario. A Monte Carlo simulation has been applied to present the probability function. Incorporated with the Monte Carlo simulation approach, this thesis proposes a well-spacing study procedure to determine the optimum coalbed methane development scenario. The study workflow is applied in a North America basin resulting in distinct Net Present Value predictions between each well-spacing design and an optimum range of well-spacing for a particular basin area.