Data aggregation is one of the challenging issues in Wireless Sensor Networks (WSNs). Traditionally, sensor nodes are assumed to be in the active state all the time, which leads to undesirable energy consumption. Moreover, most of the duty-cycle protocols are proposed to reduce energy consumption. However, the aggregation delay has been ignored during saving energy. This paper focuses on the problem of minimum-delay aggregation schedule in duty-cycled WSNs under the protocol interference model. To solve the problem, we propose an algorithm by firstly construct an aggregation tree based on the connected dominating set (CDS). Then we propose an aggregation schedule algorithm to avoid the interference, which is Leaves Schedule (LS). Different from existing works, we believe that nodes in different layers can transmit concurrently, thus it will greatly increase the number of nodes transmit concurrently and minimize the total transmission times in order to minimize the aggregation delay. Through extensive simulations, we found that the proposed scheduling method outperforms the state-of-the-art schemes.