The relevance of this study is stipulated by the necessity of designing algorithms allowing to improve the efficiency of human face detection and emotions recognition on images with complex background. Purpose: Development of algorithms and software system allowing to improve the efficiency of human face detection and in addition facial expression classification on images with complex background, in the presence of foreign objects, changing illumination, noise and different distortions. Experimental investigations to be performed into the efficiency of implemented algorithms and comparison to their existing analogs. Findings: Face detection algorithm based on Viola Jones method-is proposed to face detection on images with complex background. The model of convolutional neural network (CNN) with original structure is proposed for facial expression classification. The description of testing and training parameters, as well as comparisons with existing analogues are presented.