This work presents a framework for artificial neural network simulation and visualization using particle system concept. The whole system allows high flexibility using neural network parameters, as the thresholds to the neuron activation, during the learning and recognition phases. The framework, which consists of a library of classes, as well as a graphic interface, presents expansibility and flexibility to the development and simulation of neural network models. An application of the system to the neocognitron, a feedforward and convolutional neural network, is presented and its feasibility is shown.