Osteoarthritis is a degenerative joint disease that affects millions of people worldwide. The aims of this study were (1) to quantitatively characterise the boundary and surface features of wear particles present in the synovial fluid of patients, (2) to select key numerical parameters that describe distinctive particle features and enable osteoarthritis assessment and (3) to develop a model to assess osteoarthritis conditions using comprehensive wear debris information. Discriminant analysis was used to statistically group particles based on differences in their numerical parameters. The analysis methods agreed with the clinical osteoarthritis grades in 63%, 50% and 61% of particles for no osteoarthritis, mild osteoarthritis and severe osteoarthritis, respectively. This study has revealed particle features specific to different osteoarthritis grades and provided further understanding of the cartilage degradation process through wear particle analysis – the technique that has the potential to be developed as an objective and minimally invasive method for osteoarthritis diagnosis.