Multiresolution analysis is applied to the problem of halo identification in cosmological N-body simulations. The procedure makes use of a discrete wavelet transform known as the algorithme a trous and segmentation analysis. It has the ability to find subhalos in the dense regions of a parent halo and can discern the multiple levels of substructure expected in the hierarchical clustering scenario. As an illustration, a 500,000 particle dark matter halo is analyzed and over 600 subhalos are found. Statistical properties of the subhalo population are discussed.