The interest for constructing spatial models of electoral choice arose over half a century ago, and by the 1970s or 1980s spatial modeling had already entirely taken shape as original line of theoretical and applied-political analysis, with its own methodological basis and vectors of scientific search. Being aimed at the solution of practically applied tasks (explaining and forecasting the results of voting in concrete electoral systems), spatial modeling is constructed on a clearly explicated conceptual foundation related to essential problems of political theory, first of all the problem of electoral (and, in a broader sense, political) choice as such, its mechanisms, motivation etc. It is non-classical models of electoral choice, as well as problems and possible lines of further development of spatial modeling, that are discussed in the second part of the article (for its first part see the Polis, 2007, № 1).