The invention discloses an animal epidemic situation diagnosis method based on a Bayes network and a system, wherein the method comprises the following steps of S1, constructing an animal epidemic situation diagnosis Bayes network; S2, acquiring animal signs and environment data which comprises animal gender, age, body temperature and one-day motion number-of-steps, environment temperature and humidity of the animal, and animal epidemic situation artificial diagnosis historical record; S3, discretizing the acquired animal sign and environment data; S4, training the animal epidemic situation diagnosis Bayes network by means of discretized data; S5, calculating an animal epidemic situation occurring probability according to the trained animal epidemic situation diagnosis Bayes network; and S6, performing early-warning when the animal epidemic situation occurring probability exceeds a preset threshold. According to the method and the system, remote, automatic and accurate diagnosis and early-warning can be performed on the epidemic situation according to the sign data and the environment data of the animal, thereby effectively preventing occurrence of large-scale animal epidemic diseases, and maximally ensuring high safety of meat food.