The present study focuses on the development and application of automatic trajectory systems for swarms of unmanned aerial vehicles (UAVs) using artificial vision. Despite the growing interest and abundant literature in the field of UAVs and artificial vision, there is a significant gap in applied research, particularly in the practical implementation of these systems. A lack of experimental studies that allow direct comparisons and comprehensive measurements in physical environments is observed. The main objective of this research is to implement an automatic trajectory system using artificial vision in UAV swarms. To achieve this, a comprehensive analysis of existing literature on artificial vision strategies for object detection on routes will be carried out. Subsequently, an evasion algorithm subsystem will be developed to avoid collisions in the swarm’s trajectory, using vision models based on Deep Learning and ArUco markers used in Python programming language libraries. Experimental tests will focus on evaluating flight time and evasion capacity in a controlled environment. These tests will validate the effectiveness of the implemented system, providing practical and applied knowledge in the field of UAV swarms, addressing the need for solutions for autonomous navigation.