Manufacturing elastomer parts by injection molding with the aim of maintaining the part quality in the sense of its mechanical characteristics is a complex subject, which will be targeted in this thesis. Starting from the viewpoint that rubber compounds contain multiple components, each of those affecting the behavior of the raw material, inevitably results in natural batch variations that can occur at any time throughout the process and have a negative impact on the properties of the final product. The method examined in this work attempts to monitor the entire manufacturing cycle, verifying for the presence of potential deviations and, if existent, adjusting the process parameters so as to obtain consistent quality of the parts from cycle to cycle. For the system available, consisting of a vertical rubber injection molding machine, an industry-related multi-cavity mold containing a cold runner system, and a commercially used Styrene-Butadiene Rubber, the processing window was defined by means of a factorial design of experiment making the specified processing range. Using statistical analysis, quality functions describing this area mathematically were calculated from the test results. Based on these equations, it was possible to calculate the optimum processing point which statistically would lead to defined mechanical properties at three representative test positions on the molded parts. Due to the subsequent incorporation of measurable process response values in the calculation of the quality functions as well as in the validation experiments, it was possible to realize initial approaches that compensate emerging variations on the material, simultaneously maintaining the quality on demand. Another considerable remark of this work is that reintegration of both processing data and measurement results gives the control system a living character. This means a continuous improvement of the functions, which in return increases the accuracy of the prediction. In parallel to the experimental investigations with the injection molding machine, experiments were launched to calculate the optimum operating point by means of process simulations. The material data required for this purpose was obtained from comprehensive characterization of the rubber compounds used. Through targeted investigations, it was possible to suggest and even partly implement minor improvements in the field of simulation. Based on the overall findings, it was finally possible to realize a first prototype of a process control device. By knowing the process response values of the dosing and injection process, as well as the relevant compounding data of the crosslinking kinetics, this device calculates on-line which process adaptation has to be carried out in order to achieve a defined target value, such as a certain value of any measurable mechanical behavior of a cured rubber part. The process control method developed makes it possible to compensate for deviations and produce usable molded parts even if the crosslinking behavior of two batches of the same mixture differs significantly.