Abstract: This paper reviewed the advanced processing technique of automobile engine crankshaft rolling aggrandizement and alignment.Three key problems about straightening and alignment.of fillet rolling process was put forward.For the control of pressure fluctuation pressure in rolling process,the causes of pressure disturbance was analyzed and the internal model control method was proposed to overcome stress disturbance.A neural network modeling of internal model control system framework was constructed,the PID neural network removed the saturation was choosen and delay line identifier to proportion hydraulic system model was established.Equivalent pulse transfer function of PID neural network removed saturated properties was deduced and its inverse model was established.The application scope of the neural network was pointed out,and in order to obtain the effective data,the points to which some attention should be paid were also covered.Neural network internal model control was applied to control output pressure of proportional pressure valves of crankshaft rounded rolling machine in the hydraulic system,so the pressure fluctuations problem was solved,due to the interference caused by mechanical force of the swing of rolling dynamic clamp.