In order to overcome the limitations of traditional numerical methods in dealing with complex optimization problems and improve the efficiency of finding global optimal solutions, this paper proposes a new metaheuristic algorithm called Beaver algorithm (BA) to solve global optimization problems. According to the felling behaviour of Beavers during the construction of Beaver dams, the BA divides Beavers into three types: Inspection Beaver, development Beaver, and cutting Beaver, to simulate their group cooperation in the felling process. The development Beaver searches unknown timber land randomly to find new timber resources. The cutting Beaver rushes to the inspection Beaver to collect wood and is in the process of raiding to find wood resources. The inspection Beaver lead the whole Beaver towards the direction of the most abundant wood resources. The BA is tested on the CEC 2017 test function and compared with seven other algorithms. The results show that the BA obtains the best solution in most functions and has strong optimization ability. In addition, the BA is applied to two challenging engineering problems, such as tension/compression spring design and three-bar truss design, and is compared with the other two algorithms. The results show that the BA achieves the best optimization in engineering problems and performs better than the other two algorithms.