Abstract: For problems of current load balancing algorithms such as slow processing speed and unreasonable use of existing resources to complete task scheduling in processing data of smart mine system, a cuckoo search based weighted least connection (CS-WLC) algorithm was proposed and applied to smart mine software platform for load balancing. The algorithm synthetically considers backend server indexes including processing speed, RAM, disc IO speed, network throughput and process quantity, and calculates availability of each backend server through assigning weight to each index. Cuckoo search algorithm is used for global optimization of the backend servers according to the calculation results, so as to get an optimal solution collection. The backend servers are assigned weights considering linking number and usage frequency, and weighted least connection (WLC) algorithm is used to select a backend server with lighter load in the optimal solution collection to process real-time data and user access requests. Load balancing test is carried out on software platform of distributed fusion monitoring system. The results show that CS-WLC algorithm has shorter response time-delay and larger linking number than WLC algorithm under the condition of data increase, which verifies CS-WLC algorithm has better load balancing effect.