In order to solve the problem of time-consuming safety risk identification in safety supervision and management of electric power field operation, a safety risk identification system for electric power infrastructure construction site is proposed, which integrates front-end intelligent perception technology. In terms of hardware, the front-end sensing equipment of the Internet of Things and the UAV aircraft are designed; in terms of software, based on the front-end intelligent sensing principle, a front-end heterogeneous site construction data intelligent sensing module is established, and various information of the power infrastructure site is efficiently collected through the front-end sensing equipment of the Internet of Things carried by the UAV. The abnormal sensing data are extracted by the differential calculation method, and then the abnormal data are repaired by the genetic algorithm. Fully considering all kinds of risk factors in the electric power construction site, the construction safety risk evaluation index is determined, and combined with the fuzzy clustering maximum tree algorithm, the construction safety risk level is identified. The system test results show that the average risk identification time of the proposed system is 6.57 min, which provides more time for the safety risk prevention of on-site construction.