The Internet of things (IoT) provides a broad range of products and services to allow autonomous and secured connectivity by exchanging information between the abstract world and human activities. To ensure that these advanced systems function at their maximum capacity for a long time without any failure or fault, creating benchmarks for them is highly needed. Intelligent video surveillance system (VSS) under the Internet of video things (IoVT) paradigm is providing significant benefits being the most important safety and security tool in a variety of applications. The deterministic approaches used in most of existing VSS performance evaluation methods may often lead to either increment in cost or frequent system failures. To overcome these limitations, this paper proposes a different framework to analyze the system via the concept of reliability. The existing approaches depend solely upon the system input sequences. However, the proposed framework works on the component statistical index. Based on exponential distribution theory of reliability, the reliability of the intelligent VSS is analyzed in this paper, followed by a complete guideline to improve the current reliability of the system.