We work on the first step towards building a fiber bundle atlas on porcine hearts with diffusion tensor imaging (DTI) by (1) generating porcine heart fiber models from DTI data, (2) clustering the DTI fibers, and (3) matching cross-subject DTI fiber bundles. We also explore the effects of different distance thresholds on the DTI fiber bundle clustering and matching. Our results show that (1) complete-linkage leads to greater number of matches than single-linkage; (2) the smaller the minimum distance threshold is, the more matched bundles are obtained; (3) several fiber bundles are consistently matched across subjects with different sets of parameters.