Emotion can be impacted by a variety of environmental or ambient factors. This means, people might show different affective reactions in response to ambient factors such as noise, temperature and humidity. Annoying ambient conditions (e.g., loud noise) may negatively influence people emotion and consequently address serious mental diseases. For this, ambient factors should be monitored and managed according to the users’ preference to increase their statistician, enhance living experience quality and reduce mental-health risks. The purpose of this research is to study and predict the correlations between emotion and two ambient factors including temperature, and noise. For this, a system architecture is designed to measure user’s affect in response to the indoor ambient factors. This system is tested in three experimental scenarios each of which with 15 participants. Ambient data is collected using an IoT enabled sensor network, whereas brainwaves are collected using an EEG. The brain signals are interpreted using a well-know API to recognise emotion state. Yet, two machine learning techniques KNN and DNN are used to analyse and predict emotional statues according to changing ambient temperature and noise. According to the results, DNN has a better accuracy to predict the emotional status as compared to KNN. Moreover, it shows that both noise and temperature are positively correlated to arousal and emotional status. Moreover, the results address that noise has a greater impact on emotion as compared to temperature.