The paper deals with Statistical Process Control (SPC) applied to three original and three generated variables of copper ore upgrading by flotation. The six variables were evaluated by the SPC charts based on industrial upgrading of copper ore data gathered during one month of operation in the form of copper content in feed, concentrate and tailing. The remaining three upgrading variables were concentrate yield, copper recovery in concentrate and non-copper components recovery in tailing. Although, all variables obeyed normal distribution, considerable autocorrelation was detected between observations for all variables. For this reason, the traditional Shewhart control charts, that assume the process data generated are normally and independently distributed, resulted in many of out-of-control points which may lead to wrong decisions regarding the control of process variables. The most suitable ARIMA time series models were determined for all variables to remove autocorrelations. The ARIMA(0,1,1) model was found the best for copper content in feed, copper content in concentrate, concentrate yield and non-copper components recovery in tailing, while the AR(1) model was suitable for copper content in tailing and copper recovery in concentrate.