This thesis investigates key challenges in the design, optimization, and operation of space missions employing electric propulsion, spanning mission-level refueling strategies, propellant discovery, and plume-level modeling. First, an analytical approach is developed to assess the performance impact of on-orbit refueling for spacecraft equipped with electric propulsion. By generalizing classical performance metrics to account for refueling events, we show that evenly distributed refueling sequences maximize payload mass and that the optimal specific impulse decreases with the number of refueling stations. Applied to ESA’s SMART-1 mission, the framework demonstrates that even a limited number of refueling events can significantly increase the delivered payload and reduce specific propellant consumption. Second, we propose a machine learning pipeline for screening ionic liquids as candidate propellants for electrospray thrusters. Using molecular descriptors derived from SMILES representations and addressing severe class imbalance with hybrid resampling, we train multiple classifiers and identify a support vector machine (SVM) as the best-performing model. The SVM predicts 193 candidates from a set of more than 250,000 ionic liquids with unknown properties. Third, we investigate the chemical evolution of electrospray plumes through molecular dynamics simulations. Particle-surface collisions are modeled using a reactive force field, enabling the derivation of pseudomass spectra of collision byproducts. Results show a strong correlation between impact energy and collision byproducts, and a trend where stronger collisions produce lighter species. Intermolecular interactions within the plume are examined via simulations of monomer--neutral collisions, revealing that high-energy impacts lead to greater fragmentation, while weak interactions tend to preserve heavier species. These findings have direct implications for lifetime degradation mechanisms. Finally, a combined numerical--experimental study is conducted to characterize secondary species generated by EMI-BF$_4$ ion beams upon collision with surfaces. Simulated mass spectra are compared with experimental measurements obtained via residual gas analysis, showing good agreement and validating the molecular dynamics framework. Together, these efforts contribute to the advancement of electric propulsion systems by providing quantitative tools for system-level design, computational strategies for propellant selection, and atomistic insight into plume chemistry.