In 2017, highly fatal canine leptospirosis emerged in Sydney, Australia. Based on results of microscopic agglutination testing (MAT), serovar Copenhageni appeared to be the most common causative serovar. Prior to this, no clinical cases had been reported since 1976. In a serosurvey of healthy dogs in Australian shelters in 2004, 2.4% of 431 New South Wales dogs had serological evidence of exposure to Copenhageni, the most prevalent serovar. The aim of this study was to estimate the current prevalence of Leptospira exposure and associated serovars in healthy Sydney dogs, previously unvaccinated against Leptospira. Serum samples from 411 healthy dogs in leptospirosis hotspots and neighbouring suburbs were collected before vaccination. MAT for 23 serovars was performed at the WHO Leptospirosis Reference Laboratory in Queensland, Australia. The overall seroprevalence was 4.1% (17/411) with low titres (1/50–1/200) detected. Eleven dogs were from known leptospirosis hotspots. Eight dogs were known to hunt rodents. One dog had been in contact with a leptospirosis positive dog 1 year prior. Serovar Topaz was the most prevalent serovar (n = 5) followed by serovars Australis (n = 4), Copenhageni (n = 4), Djasiman (n = 2), Cynopteri (n = 1), Javanica (n = 1), Medanensis (n = 1), and Pomona (n = 1). In conclusion, serological evidence of exposure of dogs in Sydney to Leptospira is low, but apparently has increased since 2004. Positive titres to serovars not previously reported to cause disease in dogs could be due to low virulence of those serovars or cross‐reactivity with other serovars.