Adaptive Cruise Control (ACC) is one of the popular Advanced Driver Assistance Systems (ADAS) features, being widely implemented in many new generation vehicles. However, evaluating safety of ACC is still challenging and requires vehicles response at various scenarios. In this paper, a method for effectively evaluating ACC is presented. Effective evaluation of ACC requires underlying control architecture and for better understanding of it, two physics-based linear mathematical models are developed. These models estimate the response of ACC in a car-following scenario, i.e., predicting the longitudinal acceleration. Developed models are listed below: 1) Single Degree of Freedom Spring Damper Model (SDM). 2) Time to Collision(TTC). These developed models are fit to Naturalistic Driving Study dataset (NDS), part of Strategic Highway Research Program-2. Next, the models are fit to newer generation autonomous vehicle data which consists of vehicles possessing ACC. The results obtained from both human driver data and autonomous vehicle data are discussed in detail. Also, the parameters obtained are compared against each other and it was observed that, with the same modeling, autonomous vehicles' response has better prediction accuracy over human drivers' response.