Over the years, amplitude variation with‐offset (AVO) analysis has been used successfully to predict reservoir properties and fluid contents, in some cases allowing the spatial location of gas‐water and gas‐oil contacts. In this paper, we show that a 3-D AVO technique also can be used to characterize fractured reservoirs, allowing spatial location of crack density variations. The Cedar Hill Field in the San Juan Basin, New Mexico, produces methane from the fractured coalbeds of the Fruitland Formation. The presence of fracturing is critical to methane production because of the absence of matrix permeability in the coals. To help characterize this coalbed reservoir, a 3-D, multicomponent seismic survey was acquired in this field. In this study, prestack P‐wave amplitude data from the multicomponent data set are used to delineate zones of large Poisson's ratio contrasts (or high crack densities) in the coalbed methane reservoir, while source‐receiver azimuth sorting is used to detect preferential directions of azimuthal anisotropy caused by the fracturing system of coal. Two modeling techniques (using ray tracing and reflectivity methods) predict the effects of fractured coal‐seam zones on angle‐dependent P‐wave reflectivity. Synthetic common‐midpoint (CMP) gathers are generated for a horizontally layered earth model that uses elastic parameters derived from sonic and density log measurements. Fracture density variations in coalbeds are simulated by anisotropic modeling. The large acoustic impedance contrasts associated with the sandstone‐coal interfaces dominate the P‐wave reflectivity response. They far outweigh the effects of contrasts in anisotropic parameters for the computed models. Seismic AVO analysis of nine macrobins obtained from the 3-D volume confirms model predictions. Areas with large AVO intercepts indicate low‐velocity coals, possibly related to zones of stress relief. Areas with large AVO gradients identify coal zones of large Poisson's ratio contrasts and therefore high fracture densities in the coalbed methane reservoir. The 3-D AVO product and Poisson's variation maps combine these responses, producing a picture of the reservoir that includes its degree of fracturing and its possible stress condition. Source‐receiver azimuth sorting is used to detect preferential directions of azimuthal anisotropy caused by the fracturing system of coal.