Bluetooth is expected to be an important basic constructing component of smart homes. In a smart home environment, many devices will be portable and battery-operated, making power saving an essential issue. In this article we study the problem of managing the low-power sniff mode in Bluetooth, where a slave is allowed to be awake only periodically. One challenging problem is how to schedule each slave's sniffing period in a piconet so as to resolve the trade-off between traffic and power-saving requirements, to which we refer as the sniff-scheduling problem. We propose an adaptive protocol to dynamically adjust each slave's sniff parameters, with a goal of catching the varying, and even asymmetric, traffic patterns among the master and slaves. Compared to existing works, our work is unique. First, our scheduling considers multiple slaves simultaneously. Existing work only considers one slave, and different slaves are treated independently. Second, our scheduling is more accurate and dynamic in determining the sniff-related parameters based on slave's traffic patterns. Most work is restricted to a naive exponential adjustment in sniff interval/sniff-attempt window. Third, our proposal includes the placement of sniff-attempt periods of sniffed slaves on the time axis when multiple slaves are involved. This issue is ignored by earlier work. Extensive simulation results are presented. Among many observations, one interesting result is that with proper settings, our protocol can save significant power while achieving higher network throughput than a naive always active round-robin scheme.