Abstract: For decades, more and more experimental researches have collectively indicated that microRNA (miRNA) could play a vital role in many important biological processes and thus, it also the pathogenesis of human complex diseases. It is also noticed that the resource and time cost requirement for processing data in traditional biological method is more expensive and thus, more and more focusing have been paid to the enhancement of efective and accurate computational mechanisms for predicting potential associations between diseases. To focus towards this, researchers identified that gene is not responsible for many human diseases and instead, diseases occur due to interaction of different group of genomes that is responsible for different diseases. Hence, it is very important to analyze and associate the complete genome sequences and its associations to understand or predict various possible human diseases. To identify and predict the associations between diseases, this research work is proposed deep learning based Intelligent Human Diseases-Gene Association Prediction technique for high dimensional human diseases data sets (IHDGAP). This gene disease sequences prediction technique is proposed through deep learning method that will predict the association between the diseases. It employs Convolution Neural Network (CNN) algorithm which contains multiple number of hidden layers which is helping to predict gene patterns and its associations to predict human diseases. The proposed model, deep learning based Intelligent Human Diseases-Gene Association Prediction technique (IHDGAP) is implemented and analyzed carefully in terms of processing time, memory usage/utilization, accuracy, sensitivity, specificity and Fscore. From the experimental results, it is noticed that the proposed deep learning mechanism improves the performances of the proposed classifier in terms of accuracy, sensitivity, specificity and Fscore as compared with our previous model gene signature based Hierarchical Random Forest (G-HRF). However, it was noticed that the proposed model consumes relatively more memory and processing time as we use Convolution Neural Network (CNN) to predict gene associations.