Rapid floods usually occur in an uncertain and unpredictable time, and it will cause major disaster to environment and humanity especially at recreational areas. An IoT early warning system using Arduino technology is proposed to cater this problem. The system includes sensors for temperature, humidity, water flow, and ultrasonic measurements used for further analysis. This research aims to develop a system to monitor and detect water flow activity, such as water level and velocity, and notify of potential rapid floods earlier than estimated occurring time. The current temperature and humidity of the areas also can be recorded with the proposed device The System Development Life Cycle (SDLC) methodology was adapted for implementation of the project. Field testing at Puncak Janing Waterfall in Kedah State was chosen as test site for sensor functionality evaluation. Data is stored in a Firebase database, with an ESP32 Lo-Ra used for connectivity suitable for remote area and coding was done in Arduino IDE tool. The project successfully monitors and stores water level and velocity data and the data can be use as benchmarking the time rapid flood will occur.