The resource sharing task in the edge computing has become difficult in the recent days due to wide range of access and request in the network. The existing technique tried to reduce the burden in the edge computing to preserve the data privacy against the unauthorized access in the network during the process of resource sharing but failed to reach the expected outcome. The recurrent neural network with chaotic Henry gas solubility optimization (RNN with CHGSO) algorithm is proposed to overcome the existing problem for the unauthorized access during the resource sharing in the edge computing. The recurrent neural network (RNN) predicts the resources availability by interpreting the previous history data of edge computing. Chaotic Henry gas solubility optimization (CHGSO) algorithm finely tuned the hyperparameters of the RNN that advanced the prediction performance of the model for preserving the privacy of edge computing during resource sharing. The developed RNN with CHGSO algorithm has exhibited better results with accuracy of 0.9624, precision of 0.9143, recall of 0.9087, and f1-score of 0.9114 compared to the existing long short-term memory (LSTM) algorithm.