A theoretical model is unprecedentedly formulated for the macroscopic description of an electric field and an induced second-order nonlinearity by periodic poling in a waveguide. A fundamental field distribution between a set of two electrodes with a finite width is first proposed and then, by using linear superposition and electrode periodicity, the net electric fields generated in the waveguides are derived explicitly. A mean distribution of quasi phase matching and a generalized electrode width are introduced to analyze the poling-induced characteristics of the second-order nonlinear coefficient. Four types of poling schemes, depending on the electrode structure, are discussed from a quantitative viewpoint of calculating a quasi phase matching efficiency in the periodic poling and thus enhancing the conversion efficiency in a nonlinear optical phenomenon. The maximized efficiency and the corresponding optimized width of the shortened electrode are determined in poling experiments by means of the present approach.