Summary. Using poly(lactide-co-glycolide) (PLGA) particles for drug encapsulation and delivery has recently gained considerable popularity for a number of reasons. An advantage in one sense, but a drawback of PLGA use in another, is that drug delivery systems made of this material can provide a wide range of dissolution profiles, due to their internal structure and properties related to particles' manufacture. The advantages of enriching particulate drug design experimentation with computer models, are evident with simulations used to predict and optimize design, as well as indicate choice of best manufacturing parameters. In the present work, we seek to understand the phenomena observed for PLGA micro-and nanospheres, through Cellular Automata (CA) agent-based Monte Carlo (MC) models. Systems are studied both over large temporal scales, (capturing slow erosion of PLGA) and for various spatial configurations (capturing initial as well as dynamic morphology). The major strength of this multi-agent approach is to observe dissolution directly, by monitoring the emergent behaviour: the dissolution profile manifested, as a sphere erodes. Different problematic aspects of the modelling process are discussed in details in this paper. The models were tested on experimental data from literature, demonstrating very good performance. Quantitative discussion is provided throughout the text in order to make a demonstration of the use in practice of the proposed model.