Weather-based disease forecasts can reduce fungicide use, but uncertainty may discourage adoption. We developed a smart spray scheme (SSS) combining predicted pepper anthracnose infection risk with a 7-day nospray period after previous spray and a 21-day maximum no-spray interval. Split-plot trials were conducted in Pyeongchang, Korea, during 2021–2024, with spray scheme assigned to whole plots and cultivar to subplots. In 2021, a forecast-based spray scheme (FBS) was compared with a routine spray scheme (RSS). During 2022–2024, SSS, RSS, a minimum spray scheme (MSS), and a no-spray (NS) control were compared. Disease development was quantified using the area under the disease progress curve (AUDPC), and final incidence was analyzed using a binomial generalized linear mixed model. In 2021, FBS required six applications versus eight for RSS; neither AUDPC nor final incidence differed significantly. During 2022–2024, SSS required 6, 9, and 7 applications versus 11, 14, and 13 for RSS, respectively, reducing total applications by 42.1%. In each year, neither AUDPC nor final incidence differed significantly between SSS and RSS, whereas final incidence was significantly lower under SSS than under NS. In 2023, MSS resulted in higher final incidence than SSS and RSS. Because application timing and fungicide selection also differed between SSS and RSS, these findings reflect the performance of complete spray programs rather than the isolated effect of application number. Under the tested conditions, SSS reduced applications without a statistically detectable increase in disease relative to RSS, offering a practical framework for managing uncertainty in disease-control decisions.