The traditional material transport truck material transmission mainly relies on manual control, the disadvantage of doing so is that it is easy to cause the underload of the silo, especially in the harsh construction environment, the operator to the roof to observe the material level, which is neither safe nor accurate monitoring of the state of the material, and the manual control of the operation of the transport truck is prone to error. Therefore, this paper studies the design and implementation of automatic material transport vehicle based on visual navigation, aiming to better transport materials through the use of automatic transport vehicle based on visual navigation. Through experiments, this paper tests the improvement of the average accuracy of each operation of automatic material transport vehicle based on visual navigation. The experimental results found that the accuracy of various operations of the automatic material transport vehicle was improved by 10 % to 35% on average. From this experimental result, it can be seen that the visual navigation automatic material transport vehicle can play a good role in the material transport.