Dielectric Elastomer Generators (DEGs) are an emerging technology able to answer the need for cheap and available energy solutions, from energy scavenging at small scale to energy generation in large devices. Consisting of rubbery stretchable capacitors, DEG scan convert mechanical energy into electrical energy when their charges are displaced against the electric field during their relaxing phase. Despite having characteristics such as high energy density, low cost, and easy scalability, there are still significant challenges in their implementation. This includes the need for high voltage priming to enable the energy conversion and the adequate timing to charge/discharge the DEG. This thesis seeks to investigate and propose solutions for smart charge management in DEGs at different scales. Starting from a model-based analysis of the energy conversion phenomenon during the charging and discharging transients, we describe how undesired electrical-to-mechanical energy conversion, exhibited as an actuator-like effect, can reduce the overall energy generated during a cycle. Subsequently, we analyse different layouts of Self-Priming Circuits (SPCs), a scheme to passively promote the charge and discharge of DEGs. SPC-DEG systems can receive low voltage priming and then increase the system voltage and the amount of energy generated per cycle, eliminating the need for a high voltage time-controlled charge. Our model is experimentally validated and gives good quantitative accuracy. Using the SPC-DEG model and considering the actuator-like behaviour, we predict that the SPC-DEG systems can undergo a self-stabilising condition, which we then verify experimentally, and propose methods to estimate the emergent steady state. In addition, we develop the Self-sensing Peak Detection (SSPD) method that self-senses the deformation characteristic of DEGs and uses it as a signal to autonomously promote charge and discharge. The method was successfully implemented experimentally and has potential to optimally control DEG cycles even for unpredictable frequency and amplitude deformations. This thesis demonstrates possible solutions for charge management of DEGs and their implementation in the future.