Gob gas ventholes (GGV) are used to control methane inflows into a longwall operation by capturing it within the overlying fractured strata before it enters the work environment. Thus, it is important to understand the effects of various factors, such as drilling parameters, location of borehole, applied vacuum by exhausters and mining/panel parameters in order to be able to evaluate the performance of GGVs and to predict their effectiveness in controlling methane emissions. However, a practical model for this purpose currently does not exist. In this paper, we analyzed the total gas flow rates and methane percentages from 10 GGVs located on three adjacent panels operated in Pittsburgh coalbed in Southwestern Pennsylvania section of Northern Appalachian basin. The ventholes were drilled from different surface elevations and were located at varying distances from the start-up ends of the panels and from the tailgate entries. Exhauster pressures, casing diameters, location of longwall face and mining rates and production data were also recorded. These data were incorporated into a multilayer-perceptron (MLP) type artificial neural network (ANN) to model venthole production. The results showed that the two-hidden layer model predicted total production and the methane content of the GGVs with more than 90% accuracy. The ANN model was further used to conduct sensitivity analyses about the mean of the input variables to determine the effect of each input variable on the predicted production performance of GGVs. Published by Elsevier B.V.