The placement of a battery energy storage system (BESS) determines whether its capabilities can be effectively exploited. However, existing BESS placement studies are commonly based on a balanced network model and a single (e.g. extreme) operation scenario, while practical distribution networks are unbalanced, with complicated operation scenarios due to DG and load uncertainties. In addition, BESS placement problems are usually solved by heuristic search or mathematical programming methods. These, however, cannot balance well the efficiency and accuracy of solutions. To address these challenges, this paper proposes a two-stage hybrid optimization-based BESS placement model for unbalanced active distribution networks, considering the uncertainty of load and DG under multiple operation scenarios. Specifically, the optimal BESS capacities are decided in the first stage for minimizing costs of BESS primary investment and maintenance, while the optimal BESS charging/discharging schedules are determined in the second stage to maximize savings by loss reduction and load shifting. Then, a hybrid solution strategy of particle swarm optimization and second order cone programming is employed to solve the above BESS placement problem and thus to realize optimal and holistic calculation effect. Finally, a real Australian distribution network is simulated to verify the effectiveness and superiority of the proposed two-stage BESS placement.