Human-computer interplay has numerous crucial fields of programs and expression detection is one among them. To evaluate the facial expression, we need to analyze the variability of human faces like coloration, posture, orientation, feeling, lights and so forth. Detecting and studying facial capabilities is a pre-needful to emotion recognition. That is frequently carried out through observation of components of the face, like eyes, lips motion and many others. These are then categorized and compared to pre-defined sets of data also called a training set. At a particular in this evaluation, a person's facial expression recognition gadget is modeled exploiting the eigen face technique. The proposed technique uses Haar cascade classifier to discover the face in a photo. Fisher faces calculation has been employed for decreasing the excessive spatial property of the eigen face. The Euclidean separation between the test photograph and mean of the eigen faces is hired to anticipate emotion expressed with the aid of test face. CK/CK+ data-set is hired for training purpose. The grey scale image of the face is employed by using the version to categorize 5 simple emotions like surprise, disgust, impartial, anger and happiness.