Remote monitoring of fall actions or conditions and the everyday lifecycle of disabled losses is the vital drive of current tele-medicine. Internet of Things (IoT) and Artificial intelligence (AI) models contain deep learning (DL) and machine learning (ML) techniques, which are to be executed in the domain of medication for motorizing the recognition procedure of abnormal and unhealthy cases. A lot of applications contain real recognition of fall accidents in adult patients. Due to human gestures, it is insignificant to discover human activity with a greater level of accuracy for each application. Similarly, discovering human action was essential for automating a method to find and monitor mistrustful actions while implementing surveillance. So, this paper develops a novel Temporal Convolutional Network Based Fall Activity Recognition System for Disabled Persons (TCN-FARSDP) Algorithm in an IoT Environment. The primary intention of the TCN-FARSDP technique is to monitor and detect fall incidents for disabled persons to enhance safety and support prompt assistance. At the primary stage, the TCN-FARSDP method applies the image preprocessing stage using gaussian filtering (GF)to eliminate noise and improve the clarity of captured images. Next, the fusion of feature extraction models involves three techniques such as NASNetMobile, DenseNet121, and MobileNetV3Large are utilized to ensure a comprehensive capture of diverse image features relevant to fall activities. For the detection of fall activities, the temporal convolutional network (TCN) classifier is exploited. At the final stage, fine-tuning is performed using the Adamax to enhance the model’s convergence and stability. The performance evaluation of the TCN-FARSDP approach is implemented against a benchmark database and the experimental outcomes pointed out the better results of the TCN-FARSDP system over the existing techniques.