Targeted drug delivery with the use of nanorobots, a yet mostly theoretical but very promising future concept, is anticipated to become a significant ally in cancer treatment. The way that nanorobot systems are currently envisaged by researchers is such that they exhibit autonomous and collaborative behaviour that can be uniquely captured by multi-agent systems. In this paper, we investigate this hypothesis by describing the process of formally modelling a simple agent-based system for a simulation of targeted drug delivery. We propose a system comprising different types of nanorobots, and evaluate the effects of various parameters on the final outcome. The data that were retrieved from the corresponding simulation runs, are in support of our hypothesis, demonstrating that nanorobotic drug delivery systems can be effectively simulated by utilising intelligent agent technology.