Numerical modeling of molecular masers is necessary in order to understand their nature and diagnostic capabilities. Model construction requires elaboration of a basic description which allows computation, that is a definition of the parameter space and basic physical relations. Usually, this requires additional thorough studies that can consist of the following stages/parts: relevant molecular spectroscopy and collisional rate coefficients; conditions in and around the masing region (that part of space where population inversion is realized); geometry and size of the masing region (including the question of whether maser spots are discrete clumps or line-of-sight correlations in a much bigger region) and propagation of maser radiation. Output of the maser computer modeling can have the following forms: exploration of parameter space (where do inversions appear in particular maser transitions and their combinations, which parameter values describe a ‘typical’ source, and so on); modeling of individual sources (line flux ratios, spectra, images and their variability); analysis of the pumping mechanism; predictions (new maser transitions, correlations in variability of different maser transitions, and the like). Described schemes (constituents and hierarchy) of the model input and output are based mainly on the experience of the authors and make no claim to be dogmatic.