Background During 2020, the State of Victoria (Australia) experienced two COVID-19 waves. Both resulted in community lockdowns followed by eased restrictions. We examined variation of sexual behaviour in Victorians over time during COVID-19. Methods We conducted a repeated online survey at four timepoints corresponding with two lockdown (LD1, LD2) and two reduced restriction (RR1, RR2) periods in Victoria. A convenience sample of participants aged ≥18 years was recruited via social media and asked about their recent (past 4 weeks) sexual behaviour. Using multivariable logistic regression, we investigated variation in sexual behaviour between surveys. Results A total of 1828 surveys were completed; 72% identified as female, 69% were aged 18–29 years, 90% were metropolitan residents. The proportion reporting recent partnered sex ranged from 54.9% (LD2) to 70.2% (RR1). Across all timepoints, the most common sexual partners were regular (81.5%, n = 842) and 10.8% (n = 111) reported casual partners (e.g. hook-ups). Compared with LD1, respondents were >2-fold more likely to report casual partner(s) during reduced restrictions (RR1: aOR 2.0; 95% CI 1.1–3.7; RR2: aOR 2.8; 95% CI 1.3–5.9). Across all timepoints, 26.6% (n = 486) reported using dating apps. Compared with LD1, dating app use for face-to-face activities (e.g. dates, hook-ups) was >4-fold higher during reduced restrictions (RR1: aOR 4.3; 95% CI 2.3–8.0; RR2: aOR 4.1; 95% CI 1.9–8.8). App use for distanced activities (e.g. sexting, virtual dates) was highest during LD1 (48.8%) than other periods. Conclusions In this convenience sample, self-reported sexual behaviours fluctuated between lockdowns and reduced restrictions. While dating apps may provide a mechanism for virtual connections, this may be temporary until physical connections are possible.