A cross-sectional study was conducted on dairy farms in eight milking centres in Trinidad to determine the prevalence of risk factors for mastitis and to assess their relationship to occurrence of subclinical mastitis. The California mastitis test (CMT) was used to determine the prevalence of subclinical mastitis by estimating the somatic cell counts in bulk and composite milk. Of a total of 177 dairy farms studied, 121 (68.4%), 39 (22.0%) and 17 (9.6%) practised semi-intensive, extensive and intensive management systems, respectively. A total of 129 (72.9%), 37 (20.9%) and 11 (6.2%) farms milked cows in parlours, stanchions and pasture/out-on-field, respectively. Based on sanitary practices, 40 (22.6%), 123 (69.5%), and 14 (7.9%) farms were classified as good, fair, and poor, respectively, while 76 (42.9%) and 60 (33.9%) farms reported to rarely experience and frequently experience water shortages, respectively. Amongst the 177 farms, only seven (4.0%) used machine-milking primarily, 152 (85.9%) screened for mastitis as a routine, 18 (10.2%) teat dipped, and 49 (27.7%) practised dry cow therapy. To detect mastitis, of 152 farms involved, 20 (13.2%) used the strip cup while only two (1.3%) employed the CMT. Pipe-borne water delivered directly from the hose was the only source of water to 91 (51.4%) farms while seven (4.0%) and eight (4.5%) farms used only well and surface water (ponds and rivers), respectively. Based on bulk milk samples, the farm prevalence of subclinical mastitis was 60.5% (107 of 177) with a range from 33.3% (centre 5H) to 100.0% (centre 2B). The difference was statistically significant (P < 0.01; chi 2). However, using composite milk, the farm prevalence of subclinical mastitis was 52.5% (93 of 177) with a range from 21.2% (centre 5H) to 92.9% (centre 2B) and again, the difference in prevalence was statistically significant (P < 0.001; chi 2). Subclinical mastitis was detected in 150 (45.0%) of 333 lactating cows screened and the range of prevalence was from 17.9% (centre 5H) to 56.3% (centre 1C). The difference was statistically significant (P < 0.001; chi 2). Of a total of 14 risk factors for mastitis studied which were related to animal husbandry, personnel, mastitis control and water, only two, the herd size and practice of dry cow therapy were significantly (P < 0.05; chi 2) associated with subclinical mastitis. It was concluded that the high prevalence of subclinical mastitis in Trinidad dairy herds could significantly reduce milk production with associated economic loss. Although a majority of the risk factors studied were not found to be significantly associated with the occurrence of subclinical mastitis possibly due to confounding factors, the need to eliminate or reduce these risk factors cannot be over-emphasized.