Atypical grey prediction control system is composed of a proportional-integral-derivative controller and a grey predictor. Rather than using differential evolution algorithms to optimize the grey prediction control system as in our previous work, this study attempts to apply the gradient descent algorithm to derive the self-tuning rules for both the coefficients of proportional-integral-derivative controller and the parameters of grey model. This study also attempts to replace the original GM(1,1) with a simple equivalent GM(1,1)in order to reduce computational complexity. Analogously, an objective function in terms of the maximal overshoot, rising time, steady-state error, and settling time is used to evaluate the performance of the comparison control systems. While using the self-tuning rule, simulation results on two single-variable plants demonstrate that the grey prediction control system with equivalent GM(1,1) could perform better than the original grey prediction control system.