Apple storage quality properties(including hardness, moisture, soluble solid, total acid) were estimated through the mechanical properties of apple(including the maximum of compression, the yield force, the elastic modulus). An artificial neural network model of storage quality properties was built by the optimization algorithm of L-M(levernberg marquardt) BP neural network. The mechanical properties and the apple storage quality properties measured in the experiment were adopted as input and output to establish the BP neural network. The simulated results show that this neural network make a good estimation of apple storage quality properties through mechanical properties. When tested by five groups of Non-sample data, the relative error between the estimation of this model and the measured value is below 5%, which meets the accuracy requirement of apple storage quality properties in engineering application.