Today, the precision of micro-optics assembly is mostly limited by the accuracy of the bonding process ― and in the case of adhesive bonding by the prediction and compensation of adhesive shrinkage during curing. In this contribution, we present a novel approach to address adhesive bonding based on hybrid control system theory. In hybrid control, dynamic systems are described as "plants" which produce discrete and/or continuous outputs from given discrete and/or continuous inputs, thus yielding a hybrid state space description of the system. The task of hybrid controllers is to observe the plant and to generate a discrete and/or continuous input sequence that guides or holds the plant in a desired target state region while avoiding invalid or unwanted intermediate states. Our approach is based on a series of experiments carried out in order to analyze, define and decouple the dependencies of adhesive shrinkage on multiple parameters, such as application geometries, fixture forces and UV intensities. As some of the dependencies describe continuous effects (e.g. shrinkage from UV intensity) and other dependencies describe discrete state transitions (e.g. fixture removal during curing), the resulting model of the overall bonding process is a hybrid dynamic system in the general case. For this plant model, we then propose a concept of sampling-based parameter search as a basis to design suitable hybrid controllers, which have the potential to optimize process control for a selection of assembly steps, thus improving the repeatability of related production steps like beam-shaping optics or mounting of turning mirrors for fiber coupling.