To the problem that the reply accuracy is low and its can't answer common chat topics at the same time in agricultural man-machine dialogue, a similarity model is proposed by a comprehensive consideration of semantic words and position words, a system development process is design that agricultural problems call the interface of agricultural knowledge database, non professional issues call the interface of universal chat database. Through the test of the agricultural consulting service robot system development, the accuracy rate of the advisory reply reached 86.1%. The satisfaction rate of the system reached 80.5%.
展开▼