<p>This paper offers an approach for classifying different retinal pathologies based on a combination of type-3 fuzzy logic (T3FL) and the fractal dimension (FD). Due to the fractal nature of retinal blood vessels, the retinal FD is a logical metric to explore its use as a potential diagnostic tool. Based on these facts, a T3FL system is used to represent expert knowledge in retinal pathologies by using the FD values. In this case, T3FL helps in modeling uncertainty for enabling medical classification. Results obtained by simulation for retinal cases illustrates the potential usability of the proposal. Type-3 classification outperforms type-2 and type-1 fuzzy approaches because it can handle in a better way the uncertainty in the involved decision process. The study is, at the moment, limited to a dataset of images and simulating off-line the fractal calculation and decision process, so future work would include real-time implementation of the FD calculation and T3FL processing that should require distributed or parallel processing computing.</p>