The invention discloses an infectious disease prediction method based on a BP algorithm and an SEIR model, and relates to the technical field of infectious disease prediction methods. The method comprises the steps that 1, an SEIR model predicts the number development trend of susceptible people, infected people, latent people and rehabilitation people, historical data of prefecture and municipalepidemic situations are obtained, parameters are initialized by the SEIR model, the model is established, and a result is predicted; 2, the number of newly increased daily infected people, the total number of infected people and the number of newly increased dead people are predicted based on a neural network. According to the method, all people are specifically divided into susceptible people, infected people, latent people and rehabilitation people, SEIR model parameter initialization is carried out according to a specific scene, model establishment is based on real infectious disease data,more trend prediction better conforms to the real situation, prediction and visual graphic display are carried out for different sub-models, and the method is more visual and more convenient to use. Through the SEIR model, the index quantity development trend can be predicted, and timely decision making is facilitated.