Irregular problems require the computation of some properties for a set of elements irregularly distributed in a domain in a dynamic way. Most irregular problems satisfy a locality property because the properties of an element e depend upon the elements “close” to e. We propose a methodology to develop a highly parallel solution based on load balancing strategies that respects locality, i.e. e and most of the elements close to e are mapped onto the same processing node. We present the experimental results of the application of the methodology to the n-boby problem and to the adaptive multigrid method.