Kebakaran pada kawasan padat penduduk merupakan salah satu bencana yang berpotensi menimbulkan kerugian besar, baik berupa kerugian material maupun korban jiwa. Keterlambatan informasi dan respon penanganan menjadi faktor utama yang dapat memperbesar dampak kebakaran. Penelitian ini bertujuan untuk merancang dan membangun sistem pendeteksi dini kebakaran kawasan padat penduduk berbasis Internet of Things (IoT) menggunakan komunikasi LoRaWAN. Sistem yang dirancang terdiri dari node sensor, gateway, dan sistem monitoring. Node sensor menggunakan sensor DHT22 untuk membaca suhu dan kelembapan, sensor MQ-2 untuk mendeteksi gas/asap, serta IR Flame Sensor 5 arah untuk mendeteksi keberadaan api. Hasil pengujian menunjukkan bahwa sensor DHT22 memiliki rata-rata pengukuran suhu sebesar 36,54% dan kelembapan sebesar 49,8% dibandingkan alat ukur referensi. Sensor MQ-2 mampu mendeteksi peningkatan konsentrasi gas dan asap pada rentang 32 – 986 ppm, sedangkan IR Flame Sensor mampu mendeteksi sumber api hingga jarak maksimum 30 cm dengan sudut deteksi efektif yaitu sekitar 108°. Hasil pengujian menunjukkan bahwa Integrasi keseluruhan alat pada kondisi Non-Line of Sight (NLOS), notifikasi kebakaran berhasil diterima pada sistem monitoring hingga penghalang empat lantai dengan nilai PDR berturut-turut sebesar 99%, 100%, 98%, dan 79%, sedangkan pada penghalang lima lantai notifikasi tidak diterima karena komunikasi LoRa gagal (PDR 0%).

Fires in densely populated residential areas are among the disasters that can cause significant material losses and casualties. Delays in information delivery and emergency response are major factors that can worsen the impact of fires. This study aims to design and develop an early fire detection system for densely populated residential areas based on the Internet of Things (IoT) using LoRaWAN communication. The proposed system consists of sensor nodes, a gateway, and a monitoring system. The sensor node employs a DHT22 sensor to measure temperature and humidity, an MQ-2 sensor to detect gas and smoke, and a 5-channel IR Flame Sensor to detect the presence of fire. The test results show that the DHT22 sensor recorded average temperature and humidity measurements of 36.54°C and 49.8% RH, respectively, compared with the reference measuring instrument. The MQ-2 sensor was able to detect increases in gas and smoke concentrations within a range of 32–986 ppm, while the IR Flame Sensor successfully detected fire sources at a maximum distance of 30 cm with an effective detection angle of approximately 108°. The overall system integration test under Non-Line of Sight (NLOS) conditions demonstrated that fire notifications were successfully received by the monitoring system through obstacles of up to four floors, with Packet Delivery Ratio (PDR) values of 99%, 100%, 98%, and 79%, respectively. However, communication failed when obstructed by five floors, resulting in no fire notifications being received (PDR 0%).