The objective of the study was the investigation of the behaviour of cathelicidin-1 in the milk after experimental infection with two prominent bacterial pathogens (experiment 1: Mannheimia haemolytica, experiment 2: M. haemolytica and Staphylococcus chromogenes) as a potential early indicator for diagnosis of mastitis in sheep. In two experiments, after bacterial inoculation into the udder of ewes, bacteriological and cytological examinations of milk samples as well as proteomics examinations [two-dimensional gel electrophoresis analysis (2-DE) and matrix-assisted laser desorption/ionization time-of-flight mass spectrometer (MALDI-TOF MS) analysis] were performed sequentially. Cathelicidin-1 was detected and spot densities obtained from PDQuest v.8.0 were recorded. Associations were calculated between cell content and spot densities as well as between presence of mastitis in a mammary gland at a given time-point and detection of cathelicidin-1 in the respective milk sample. All inoculated mammary glands developed mastitis, confirmed by the consistent bacterial isolation from mammary secretion and increased leucocyte content therein. Spot density of cathelicidin-1 in samples from inoculated glands increased 3 h post-inoculation; spot density of cathelicidin-1 in samples from inoculated glands was higher than in samples from uninoculated controls. There was clear evidence of correlation between cell content and cathelicidin-1 spot densities in milk samples. There was significant association between presence of mastitis in the mammary gland and detection of cathelicidin-1 in the respective milk sample; overall accuracy was 0.818—this was significantly greater during the first 24 h post-challenge (0.903) than after the first day (0.704). In conclusion, detection of cathelicidin-1 in milk was significantly associated with presence of mastitis in ewes. The associations were stronger during the first 24 h post-infection than after the first day. Cathelicidin-1 has the advantage that it can be a non-specific biomarker, as simply a “positive”/“negative” assessment would be sufficient.