Sensor-based healthcare development takes a new height in recent days over traditional ways of treatment in healthcare [1]. Smart healthcare system involves various participants, including medical practitioners, medicinal experts, sanatoriums, medical centers, therapeutics, and research organizations [2]. In order to find the most reliable and most trustworthy way for the early predictions of chronic disease or life-threatening disease, sensor-based approaches play a vital role in gathering the huge flow of data and make it obtainable for the analysis purpose. The sensor-based approach has numerous dimensions, covering medical research, assessment, making decisions, management of healthcare, and early chronic disease prevention. Our findings highlight result-oriented wearable sensors, implantable devices, and ambient monitoring systems, as well as various method based on machine learning and their applications in early disease prediction, geriatric care, chronic disease prediction, and preventive medicine. We also identify significant research gaps and suggest future initiatives to improve the efficacy and uptake of sensor-based healthcare solutions. The technique of predicting chronic illnesses early is challenging and requires extensive knowledge as well as experience [3]. Healthcare systems have adopted Internet-of-Things-based technology to gather sensor data for diagnostic evaluation, but the precision and reliability of these evaluations still remained a complex challenge. The wearable framework employs a variety of categorical sensor that accomplish a wider range of targets, including activity monitoring, finding disorder, safety observation, rehabilitation at home, real-time health monitoring, and treatment evaluation [4]. Advancements in technological approaches, such as the incorporation of various machine learning algorithms with medical sensors, are expected to improve the precision and effectiveness of disease evaluations, as well as open up opportunities for clinical decision support and early detection of most commonly chronic diseases such as cardiac disease, heart attack, respiratory disease, diabetes, Parkinson’s disease, neurodegenerative disease, and cancers. Some machine learning methods include support vector machine, RNN, convolutional neural network, recurrent neural network, CNN-BiGRU, bidirectional LSTM, MDCNN, decision tree, and Naive Bayes that show good results [5].