Fires cause significant damage and destruction to many individuals every year globally, necessitating improved approaches to fire detection and classification. This paper presents a multi-sensor system that detects and classifies fire using a camera and several gas and temperature sensors based on the nature of the burning material. It also dispatches notifications about the fire source and determines and executes an appropriate response to the detected fire. The study investigated three machine learning models (Neural Networks, Random Forest, and K-Nearest Neighbours) and three pre-trained Convolutional Neural Network models (VGG16, MobileNetV2, and EfficientNetBO) for sensor and image data classification, respectively, in indoor fire detection. Random Forest and VGG16 were the best­ performing models for sensor and image classification tasks. The results obtained show that the sensor data classifier obtained an accuracy of 99.35%, while the image classifier had an accuracy of 95.83%, with identical values for precision and recall. The overall system response time was measured to be approximately 15 s on average, demonstrating a fairly quick and accurate response in determining the nature of fires in indoor environments.