The aim of this study is to predict the heading day of rice by using machine learning. The heading day denotes the blooming day in the rice cultivation, which is one of the important days in the rice cultivation stage. The number of days from the planting day to the heading day is taken as the objective variable. The cultivation variables such as the weather data, the position of the field and so on are taken as the explanatory variables. The prediction model is defined by using Xgboost algorithm. The prediction model showed that the average error is 3.19 days.