Frailty assessment is a critical approach to assessing the health status of older people. There is interest in the nature and degree of frailty in various physical, environmental, social, psychological, and physiological domains. The physical domain has proved to be the most important as the degree of frailty can be assessed objectively. Various clinical tools deployed by geriatricians in assessing frailty can be grouped into two categories of questionnaire-based and physical performance analysis. In the former approach, answers vary from double (yes or no) to multiple choices and are marked accordingly. Whereas, in performance analysis, the time taken by a subject to complete a physical task such as walking for 3 m is measured. The questionnaire-based method is subjective, and the time-based performance analysis does not identify the kinematic characteristics of motion and its root causes of it. However, motion characteristics are crucial in measuring the degree of frailty. Overall, reliable and objective frailty assessment tool that can be adapted universally is unavailable. This study investigated the development of a quantitative, objective, non-invasive frailty measurement method that can assess frailty in older people. The frailty was assessed by measuring the body motion characteristic using an array of wearable inertial sensors embedded in a motion capture system and advanced machine learning methods. A set of experimental procedures was designed to assess the frailty in older people by performing various tasks such as “Pick-up an object from the floor,” “Sit to Stand,” and “Stand to Sit.” The data obtained through these activities were analysed, and features were extracted from various sensor data to create datasets for training algorithms. Artificial Neural Network, Decision Tree, Random Forest, Support Vector Machine, Deep learning Model, and Gaussian Mixture Model were applied to the data set to assess and determine the frailty objectively. This study focused on a quantitative assessment of frailty without taking any age-related, cognitive, and pathological factors into consideration, as it is beyond the scope of this study. The results of developed algorithms were benchmarked against conventional clinical measuring tool, i.e., Rockwood Frailty Scale. The analysis of the frailty assessment by the proposed algorithms showed that these proposed methods could serve as an alternative objective frailty assessment tool to the existing subjective frailty measuring methods. The performance of various classifiers showed a significant difference between the non-frail and frail classes.