The digital proportion control is introduced to improve the performance of the analog adaptive interference cancellation system (ICS). For the high frequency parts of the signals after multiplier are not required,the sampling frequency need not satisfy the sampling theorem for high frequency. Because the sampling,calculation and output expend time in digital control,the ideal condition,delay condition and delay-wait condition are taken into account. Through analyzing the system model with three conditions,we gain the stable conditions of the system,the optimization step factors that can make the system converge fastest and the formulas of the interference cancellation ratios (ICRs). One step convergence can be accomplished under ideal condition,whereas the system can not converge in one step under delay condition and delay-wait condition. The calculation results show the convergence speed of delay-wait condition is slower than that of delay condition. The ICR is improved with the increase of the step factor which is in stable bound,but the convergence speed is decreased if the step factor exceeds the optimization step factor. In order to avoid that confine,the method of amending the steady state weight to improve the ICR is proposed. The analyses are in agreement with the computer simulations. The digital proportion control is introduced to improve the performance of the analog adaptive interference cancellation system (ICS). For the high frequency parts of the signals after multiplier are not required, the sampling frequency need not satisfy the sampling theorem for high frequency. Because the sampling, calculation and output expend time in digital control, the ideal condition, delay condition and delay-wait condition are the system, the gain step stable of the system can make the system converge fastest and the formulas of the interference cancellation ratios (ICRs). One step convergence can be accomplished under ideal condition, while the system can not converge in one step under delay condition and delay-wait condition. the convergence speed of delay-wait condition is slower than that of delay condition. The ICR is improved with the incre ase of the step factor which is in stable bound, but the convergence speed is decreased if the step factor exceeds the optimization step factor. In order to avoid that confine, the method of amending the steady state weight to improve the ICR is proposed. analyses are in agreement with the computer simulations.