Bovine mastitis is the most costly disease for dairy farmers, hence, control measures to prevent it are crucial for dairy farm sustainability. Staphylococcus aureus is considered a major mastitis pathogen because of its impact on milk quality and low cure rates. Prevention of S. aureus mastitis includes segregation of infected animals, whilst treatment of such animals should be performed for a longer time to improve cure rates. This makes identification of S. aureus infected quarters and animals of significant importance. The experiments reported in this research paper aimed to develop and validate a sensitive method for magnetic detection of S. aureus and of the Staphylococcus genus in raw milk samples. Mastitic milk samples were collected aseptically from 47 cows with subclinical mastitis, from 12 Portuguese dairy farms. Forty nine quarter milk samples were selected based on bacteriological results. All samples were submitted to PCR analysis. In parallel, these milk samples were mixed with a solution combining specific antibodies and magnetic nanoparticles, to be analysed using a lab-on-a-chip magnetoresistive cytometer, with microfluidic sample handling. The antibodies used in this work were a rabbit polyclonal IgG anti-S. aureus ScpA protein and a mouse monoclonal IgM anti-S. aureus ATCC 29740. This paper describes the methodology used for magnetic detection of bacteria, including analysis of false positive/negative results. This immunological recognition was able to detect bacterial presence above 100 cfu/ml, independently of antibody and targeted bacteria used in this work. Comparison with PCR results showed sensitivities of 57·1 and 79·3%, specificity values of 75 and 50%, and PPV values of 40 and 95·8% for magnetic identification of Staphylococci species with an anti-S. aureus antibody and an anti-Staphylococcus spp. antibody, respectively. Some constraints are described as well as the method's limitations in bacterial quantification. Sensitivities and specificities require to be improved, nevertheless, the methodology described may form the basis for a means of identifying S. aureus infected cows at the point of care.