This paper proposes a peculiar and very important developing area concerns the remote monitoring of elderly or ill people. Indeed, due to the increasing aged population, Human-Computer Intelligent Interaction (HCII) systems able to help live independently are regarded as useful tools. In this context recognizing people emotional state and giving a suitable feedback may play a crucial role. The purpose of speech emotion recognition system is to automatically classify speaker's utterances into seven emotional states including anger, boredom, disgust, fear, happiness, sadness and neutral state. Emotions have been classified separately for male and female based on the fact male and female voice has altogether different range. It provides a solution by improving interaction among human and computers, thus allowing human-computer intelligent interaction. The system is composed of two subsystems: 1) gender recognition (GR) and 2) emotion recognition (ER). It distinguishes a single emotion versus all other possible ones as in proposed numerical results. Speech based emotion recognition system consists of four principle parts: Feature Extraction, Feature Selection, Database and Classification. Nowadays, the research is focused on finding powerful combinations of classifiers that increases the classification efficiency in real-life speech emotion recognition applications. From these acoustic signals, this project will calculate pitch, short time energy, zero crossing rate and Mel frequency cepstral coefficients, and correlate it to emotions of the driver. We also define these features and the feature extraction methods. In paper, a demonstration on how one can distinguish the emotion based on these features (or combination of features) by testing them over Berlin emotion database