The multi-index evaluation model of agriculture machine navigation system was proposed, and the cost, accuracy, adaptability, stability and computing complexity were used as five indexes which mainly affect the navigation system performance. The weight of each index was decided by combining AHP and entropy method, which has the advantages of subjective and objective. The RBF neural network was used to build the evaluation model. Data analyses showed that the method was easy to use, and overcomed the influence of subjective factors.