A cross-sectional survey was conducted to determine the seroprevalence of Leptospira in a cohort of horses and to evaluate potential risk factors for Leptospira seropositivity in horses in New Zealand. The convenience sample included 499 Thoroughbred racing and breeding horses from 25 commercial properties in North Island, New Zealand. A questionnaire was used to collect demographic data on horses and property-level information on grazing and management practices, pest (rodent) management, access to natural waterways, other livestock on the property, and possible contact with wildlife. The microscopic agglutination test was used to test sera for serovars Ballum, Copenhageni, Hardjo (bovis), Pomona, and Tarassovi. Logistic regression was used to investigate the risk factors for Leptospira seropositivity to at least one serovar and for each serovar individually. A total of 124 (25%, 95% confidence interval (CI) 21–29%) horses had positive titres to any one of the five serovars. The seroprevalence of Ballum, Copenhageni, Hardjo (bovis), Pomona, and Tarassovi was 5% (95% CI 3–7%), 9% (95% CI 7–12%), 6% (95% CI 4–8%), 6% (95% CI 4–8%), and 6% (95% CI 4–8%), respectively. Broodmares, compared to racehorses and alternately grazing horses with sheep, increased the odds of exposure to any one serovar, whilst grazing the same time as sheep and alternately grazing horses with cattle increased the odds of exposure to Ballum and Hardjo (bovis), respectively. Historical exposure to Leptospira in racing and breeding horses was identified, and risk factors were consistent with pasture-based exposure.