We introduce a numerical technique to investigate the temperature distribution in arbitrarily complex plasmonic systems subject to external illumination. We perform both electromagnetic and thermodynamic calculations based upon a time-efficient boundary element method. Two kinds of plasmonic systems are investigated in order to illustrate the potential of such a technique. First, we focus on individual particles with various morphologies. In analogy with electrostatics, we introduce the concept of thermal capacitance. This geometry-dependent quantity allows us to assess the temperature increase inside a plasmonic particle from the sole knowledge of its absorption cross section. We present universal thermal-capacitance curves for ellipsoids, rods, disks, and rings. Additionally, we investigate assemblies of nanoparticles in close proximity and show that, despite its diffusive nature, the temperature distribution can be made highly non-uniform even at the nanoscale using plasmonic systems. A significant degree of nanoscale control over the individual temperatures of neighboring particles is demonstrated, depending on the external light wavelength and direction of incidence. We illustrate this concept with simulations of gold sphere dimers and chains in water. Our work opens new possibilities for selectively controlling processes such as local melting for dynamic patterning of textured materials, chemical and metabolic thermal activation, and heat delivery for producing mechanical motion with spatial precision in the nanoscale.