Facial image emotion recognition detection is an increasingly important area of research in computer vision and artificial intelligence. The development of deep learning models has provided significant improvement accuracy and reliability of the system. Model was trained with a total 2,489,095 parameters, with 35,887 images with 100 epoches. The deep learning model applied Conv.Relu features extraction and Artificial Neural Network (ANN) techniques Convolutional Neural Network (CN)  classification, the model generated an accuracy of 76.83% on the training datatset and 65.38% on the validation dataset, the model was performs well but it still overfifed as at training accuracy was at 65%.Seven emotions such happy, sad, neutral, disgust, surprise and fear where evaluated using precision, recall and f-score of 0.26 while the disgust emotion has the worst precision, recall and f-score of 0.01.