The emotion of the children with Autism Spectrum Disorder (ASD) cannot be identified and recognized easily. The research in automated emotion recognition methods is steadily growing thrust in the last few years due to applicability in various domains which would benefit from a clear understanding of human emotional states. The studies have shown that a human’s physiological parameters are directly related to his/her psychological reaction from which the emotions can be estimated. There is a strong relationship between human emotion and physiological signals. The major aim of this work is to identify preferable Artificial Intelligent Ensemble Feature Selection (AIEFS) framework and Heterogeneous Ensemble Classification (HEC) model for such a concept. The experiment was necessary to achieve the uniformity in the various aspects of emotion elicitation, data processing, feature selection using EFS, and estimation evaluation using HEC and in order to avoid inconsistency problems. Here, three base classifiers such as Support Vector Regression with Genetic Algorithm (SVR-GA), Multinomial NaiveBayes (MNB) and Ensemble Online Sequential Extreme Learning Machine (EOS-ELM) that learn different aspects of the emotion dataset samples are used together to make collective decisions in order to enhance performance of health-related message classification. The results indicate that the combination of AIEFS with HEC exhibited the highest accuracy in discrete emotion classification based on physiological features calculated from the parameters like ECG, respiration, skin conductance and skin temperature. Specific discrete emotions were targeted with stimuli from the IAPS database. This work presents experiment based comparative study of four feature selection methods and five machine learning methods commonly used for emotion estimation based on physiological signals.