The cloud model is a powerful tool to achieve qualitative and quantitative uncertainty transforming. It integrates the fuzziness and randomness of objective things and solves the problem of nonlinear and uncertainties. For the nonlinearity, time-variability and uncertainty of excitation system in ship power system, the cloud model is applied to synchronous generator excitation control and two dimensional cloud model of PID excitation is designed. The forward cloud model is utilized to represent the voltage difference of synchronous generator and rate of voltage difference change concepts. Then the backward cloud generator is employed to generate the three parameters of PID concepts. Thus the cloud reasoning rule of double condition multi-rules of two dimensional cloud model is formulated and PID parameters of excitation controller are self-tuned. Through the establishment of models of synchronous generator controllable phase compound brushless excitation system and the two dimensional cloud model PID excitation system, the result of simulation shows that the two dimensional cloud model PID excitation controllable system model has better robustness and better adaptive ability.