In order to realize the intelligent diagnosis of potato early blight, with the combination of the linear superposition of quantum computing ideas and adaptive neural network computation, a quantum neural network model for diagnosis of potato early blight was built. The model used multiple quantum energy levels of the hidden layer activation function of the linear superposition of quantum neuron model. Fuzzy decision of disease diagnosis was effectively solved. Uncertainty characteristics of sample data was adaptively given in training process to determine. The algorithm overcame the disadvantages of local minimum and increased learning efficiency and training speed. The simulation results showed that the diagnosis accuracy reached to 96.5%. 