Coronavirus is an extremely infectious and fatal disease that can be spread straight from one individual to another. COVID-19 is currently causing a lot of concern worldwide since it is difficult to detect and prevent. The Internet of Things (IoT) coupled with the wireless sensor network (WSN) has an impact on lowering the medical expenses and improving the treatment results of the infected individual. This paper proposes a secure and energy-efficient WSN architecture combined with machine learning and IoT to recognize and observe the Covid-19 patients. The proposed architecture is designed to determine whether a person has COVID-19 or a typical cold, depending on their symptoms. The proposed architecture utilizes the supervised machine learning techniques such as random forest classifier, multi-layer perceptron, Naive Bayes, logistic regression, support vector machine classifiers to improve the precision of COVID-19 investigation. Energy efficiency is a significant obstacle for the sensor devices' longterm sustainability because signal transmission from many biosensors to the cloud consumes a large amount of energy. The proposed architecture substantially enhances the WSN's power efficiency as well as its longevity. The findings show that the identified variables can assist in forecasting the probability of having a more serious illness in COVID-19 patients and can aid with health resource allocation. © 2021 Elsevier B.V., All rights reserved.