In today’s modern technology the thermal camera plays a major role in all security applications, surveillance applications, etc. And, in pandemic situations the thermal camera plays a major role, because of its nature of picture the scene based on the thermal radiation emitted by the objects in a scene. Hence, regardless of any weather conditions and lighting conditions, an object can be visualized. So, in general, there is huge demand for effective object classification system in thermal videos. The proposed system uses Dynamic STP (Spatial-Temperature Pattern) & TTP (Temporal-Temperature Pattern) based FCNN (Fully Convolutional Neural Network) for object classification in thermal videos. There are two novel methods used in the proposed system. First one is Temperature histogram-based object detection and another one is Dynamic STP & TTP based FCNN for object classification. The nature of FCNN cannot be fully utilized in thermal videos due to the lack of texture information. The dynamic STP & TTP is designed for each object as a kernel to extract the STP, TTP feature map of thermal videos. The fully connected layer with soft max classification is used for classification of objects. The experimental result shows that the proposed system outperforms the state of art methods.