During the last two decades, the production of renewable energy by anaerobic digestion (AD) in biogas plants has become increasingly popular due to its applicability to a great variety of organic material from energy crops and animal waste to the organic fraction of Municipal Solid Waste (MSW), and to the relative simplicity of AD plant designs. Thus, a whole new biogas market emerged in Europe, which is strongly supported by European and national funding and remuneration schemes. Nevertheless, stable and efficient operation and control of biogas plants can be challenging, due to the high complexity of the biochemical AD process, varying substrate quality and a lack of reliable online instrumentation. In addition, governmental support for biogas plants will decrease in the long run and the substrate market will become highly competitive. The principal aim of the research presented in this thesis is to achieve a substantial improvement in the operation of biogas plants. At first, a methodology for substrate inflow optimization of full-scale biogas plants is developed based on commonly measured process variables and using dynamic simulation models as well as computational intelligence (CI) methods. This methodology which is appliquable to a broad range of different biogas plants is then followed by an evaluation of existing online instrumentation for biogas plants and the development of a novel UV/vis spectroscopic online measurement system for volatile fatty acids. This new measurement system, which uses powerful machine learning techniques, provides a substantial improvement in online process monitoring for biogas plants. The methodologies developed and results achieved in the areas of simulation and optimization were validated at a full-scale agricultural biogas plant showing that global optimization of the substrate inflow based on dynamic simulation models is able to improve the yearly profit of a biogas plant by up to 70%. Furthermore, the validation of the newly developed online measurement for VFA concentration at an industrial biogas plant showed that a measurement accuracy of 88% is possible using UV/vis spectroscopic probes.