Abstract. This article presents the development of an anti-drone system based on artificial intelligence (AI) for detection, radio-frequency jamming, and destruction of unmanned aerial vehicles (UAVs). Drawing on the analysis of UAV use in modern conflicts (Syria, Ukraine) and national regulations of the Republic of Kazakhstan, system requirements were defined, taking into account real sensor parameters (X-band radar with RCS 0.01 m², IR sensor sensitivity 0.1 K), jamming power (50 W in the 400–6000 MHz range), operational conditions (+50 °C, dust storms), and external factors (weather reducing probabilities by 20%). A probabilistic model with AI coordination (neuro-symbolic approach) was proposed, providing novelty through adaptation to Central Asian conditions. The simulation, implemented using Monte Carlo with 1000 iterations, is published in an open GitHub repository with replication instructions. Modeling shows validated results: detection probability of 95.8%, jamming effectiveness of 54–78%, and destruction probability of 70.7–84.6% for guided and autonomous drones at distances of 5–8 km. Comparative analysis with analogs (“Drone Dome”, “Pantsir-S1”) demonstrates superiority in range (10 km vs. 3.5 km), cost efficiency, and adaptability. Computational complexity analysis (O (1) per drone) and optimization pathways (ML for trajectory prediction, distributed data processing) confirm practical applicability. The system is expected to strengthen Kazakhstan’s defense capacity and reduce dependency on foreign technologies.

Keywords: anti-drone complex, artificial intelligence, probabilistic model, RF suppression, verification, defense capability.