This thesis explores the analysis and design of model predictive flocking for the distributed control of multi-agent systems. A novel framework allowing for an independent tuning of attractive and repulsive inter-agent forces is introduced. With respect to the analysis, challenges and shortcomings in existing results are discussed, and a line of reasoning based on the stability of sets of equilibria is presented. Additionally, a novel framework for flocking with elliptical desired formations is introduced for predictive and non-predictive settings, providing analysis results for both cases. The proposed frameworks are validated through simulation.