The input and output characteristic curves of a transistor are such that they can be modeled by spline functions. The problems of such a representation are two-fold. First, the best representation for the spline function must be determined, and second, all the data needed to implement the spline functions must be obtainable from easily measured parameters of the transistor. In this work, the natural Hermite cubic spline has been found to provide the needed results. The data needed to implement the Hermite cubic spline include its knots, the values at the knots, and the slopes at the knots, each of which have been defined in terms of measurable transistor parameters. The ensuing spline model has been implemented on the computer and characteristic curves of different transistors computed by the spline model have been compared with the actual curves. Results of this comparison are shown as well as the results of a comparison of the spline model with the Ebers-Moll transistor model.