Fuzzy information of corn juices was automatically evaluated based on a sensor array. The concept of weights was introduced for different taste sensory evaluation aspects of corn juices. The conversion of the qualitative and quantitative information was achieved. At the same time, adjustment of a comprehensive cloud model was completed based on the difference of weight. According to the requirements of the different evaluation aspects of corn juices including sweetness, soursweet and flavor, sensor array were analyzed and combination of different sensor array signals were obtained. Fuzzy neural networks were built for prediction of corn juices taste fuzzy information. The information for different aspects collected from sensor array was input. The information from cloud model according to sensory evaluation was output. With training fuzzy neural network, fuzzy layer center value, the fuzzification layer node width values and fuzzy decision-making regulation parameters were obtained to determine the network structure. The forecast analysis showed that the system allowed good effect with 0.00243~0.09177 error rate in the process of automation evaluation of fuzzy information for corn juices.