Appropriate and accurate detection of the people like the elderly and patients suffering from imbalanced-oriented problems is a significant step to implement the necessary actions in order to enhance quality of life and increase their independence during their life. The aim of this paper was to investigate the differences in performance of parametric and non-parametric classifiers in recognition of daily activities of people. As daily affairs of human being can change dynamically, new patterns may emerge as well as an old habit may be forgotten or stopped to be used. It is difficult to build a classifier that will have a full description of all possible ways to perform an activity. In this research, for exact identification of everyday activities, data collected from information of four accelerometers in noteworthy points of body which have the greatest role in the process of identifying and separating activities. This dataset was acquired from UCI Machine Learning database. After extraction of convenient features, different algorithms including KNN and Neuro-fuzzy classifiers were applied to data to check the rate of accuracy in exact recognition of daily activities. The results showed that the KNN classifier as anon-parametric classifier encompasses the highest accuracy rate of about 99.758[Formula: see text]. It is concluded that recognizing activities using multiple small body-worn accelerometers and implementing an optimized unobtrusive approach can lead us to a significant result.