This thesis is concerned with the problem of fluctuating mass and volume streams in waste treatment plants. The heterogeneous material from the mixed municipal waste creates a fluctuating output despite the continuous feeding of the hopper. The suggested solution is to utilize a control loop on the shredder that minimizes output fluctuations and optimizes the use of the conveyor belt and following aggregates in the waste treatment plant. The experimental control loop showed no significant improvement. However, using time series analysis, a prediction model was developed. A different version of equally spaced data points is used for this model, the mass series. This prediction model is up to 75 % better than the defined benchmark of using the current value to predict the next values. The data was taken from the large-scale industrial experiments performed with mixed municipal waste as a part of the ReWaste F Project.