This research is to develop the statistical process control system involving a nonstationary linear stochastic process. State space model is used to describe characteristics and properties of the process. To estimate and forecast a change in the process, the filter is constructed by assuming the noise terms are multivariate generalized Laplace distributed. This filter possesses a robust property to outliers in the noise terms such that a magnitude of process change can be approximated precisely. Including an adaptive factor, an effectiveness of the robust filter should be improved. An investigation of filter performance is established by Monte Carlo simulation. The results show that the robust filter with multi adaptive factors performs well in terms of both a mean squared state error and a convergence rate of mean squared state error. Statistical process control system is introduced as an integration of a process adjustment in a conception of engineering process control and a control charting technique as a tool in statistical process control. To enchance an effectiveness of the statistical process control system, the robust filter should be applied. Process adjustment is obtained by linear stochastic control theory as the noise terms assumingly have multivariate generalized Laplace distributions. Then, the control input variable as a function of the robust state variable is provided to bring the process back to a target. However, the process adjustment does not ask to search for causes of process changes. Dealing with no elimination of causes, those changes may be occurred repeatedly. Thus, the control chart of forecast error obtained by the robust filter is constructed to monitor a process and to point out a time period when a process change is occurred. The results show that statistical process control system yields a reduction in process variation and provides an absolute solution to maintain a process in a state of statistical control.