Rice blast caused by Pyricularia grisea is the most destructive disease in rice production in Japan. To control this, rice varieties with new complete resistance genes were bred and released. However, the resistances were broken down within a few years because of the emergence of compatible blast races. To prevent the breakdown, a mixture of near-isogenic lines with different complete resistance genes to blast was first introduced for multilines in 1995 in Japan. For effective control of blast, the author aimed to clarify the degree of blast suppression using multilines. This study was conducted to investigate the mechanisms of blast suppression and was developed as a simulation model for forecasting leaf blast epidemics to analyze the degree of leaf blast suppression in multilines. The autoinfection ratio was a major factor influencing the degree of leaf blast suppression in multilines. A simulation model for rice blast epidemics in multilines was developed, and the parameter of autoinfection was introduced in the model. The calculated data using the model fit well with the observed data of leaf blast development. Parameters of induced resistance, mutation rate and stabilizing selection were also introduced to clarify the mechanisms of leaf blast suppression. In addition, distribution of blast races in multilines was clarified, and a high-throughput method for race differentiation was developed.