Liquid/liquid dispersions, which are mixtures of (nearly) immiscible liquids, are an integral part of many technical processes and everyday consumer products. For example, in technical processes the interfacial area is increased by dispersion to enhance mass and heat transfer; or consumer products like lotions and creams consist of stabilised oil/water dispersions. The process efficiency and resulting product quality are particularly influenced by the drop size distribution that consequently is the essential parameter of liquid/liquid dispersions. The drop size distribution is determined by the fundamental phenomena of drop breakage and coalescence which are still not understood in detail and are in the focus of current research up to now. Especially, the complex interactions at different scales during coalescence of two drops still prevent a consistent description of the coalescence process. Available modelling approaches consider mechanistic steps during coalescence but need stochastic methods to describe the coalescence probability within a liquid/liquid system. Due to various superimposing influencing factors on coalescence, available models do not implement these influences consistently and predict partly contradictory dependencies. Consequently, it is mandatory to evaluate and validate existing coalescence descriptions and develop reliable models. In this thesis a systematic approach of coalescence research was pursued: single drop experiments were used to determine fundamental influencing parameters and to evaluate and develop coalescence models. To establish a solid data base of experimental results, an automated test cell was established using high speed imaging. With this set-up the coalescence process of two colliding drops was examined under defined conditions. An automated image analysis was developed to process the high amount of gained data. Using detailed investigations of the influences of drop size and drop collision velocity, available coalescence models were evaluated and validated. Furthermore, it was possible to determine the numerical parameter of coalescence efficiency models. Due to the systematic single drop experiments, the coalescence parameter was determined independently from drop breakage. Based on these findings a transfer to the description of drop swarms in dynamic liquid/liquid systems was performed using population balance equations as modelling framework. In this framework the fundamental phenomena of drop breakage and coalescence are implemented separately and can be modelled independently. Applying the determined coalescence parameter, the simulations were validated by drop size distribution measurements in a stirred vessel. Moreover, a coalescence efficiency model was developed which implements electrostatic repulsions due to ion partition effects at the interface in liquid/liquid systems. This model was applied to describe an observed coalescence inhibition at high pH values in a stirred tank.