Currently, surveillance system and fire detector and surveillance system are developed separately. This paper focuses on development of smart surveillance system with fire prevention function. The existing fire detectors use smoke sensor, heat sensor that can just detect fire during growth stage. In this paper, RGB and CIE L*a*b algorithm are used to detect fire spark or flame using embedded camera. This system is developed to reduce loss of fire incident by detecting early stage fire (fire spark/flame) and notify the early stage to community, thus fast action can be taken. The developed system can be remotely controlled from anywhere that under IEEE standard connectivity. User can control the system from smartphone using Telegram application. The system can be installed in various type of premises such as factory, schools, private building and welfare center. The user will receive video and picture once the fire spark is detected. This module is important to handle false alarm case, where user can identify the incident using the received video/picture. Once confirmed, the user could decide to stop sirens if false alarm or notify authorized department for further action. These smart functions could help in reducing large damage since the authorized can be notified soon as fire spark occurs. The design, development and testing process will be discussed in this paper.Currently, surveillance system and fire detector and surveillance system are developed separately. This paper focuses on development of smart surveillance system with fire prevention function. The existing fire detectors use smoke sensor, heat sensor that can just detect fire during growth stage. In this paper, RGB and CIE L*a*b algorithm are used to detect fire spark or flame using embedded camera. This system is developed to reduce loss of fire incident by detecting early stage fire (fire spark/flame) and notify the early stage to community, thus fast action can be taken. The developed system can be remotely controlled from anywhere that under IEEE standard connectivity. User can control the system from smartphone using Telegram application. The system can be installed in various type of premises such as factory, schools, private building and welfare center. The user will receive video and picture once the fire spark is detected. This module is important to handle false alarm case, where user can identify the incident using the received video/picture. Once confirmed, the user could decide to stop sirens if false alarm or notify authorized department for further action. These smart functions could help in reducing large damage since the authorized can be notified soon as fire spark occurs. The design, development and testing process will be discussed in this paper