In order to improve the transmission performance of Bluetooth system in dynamic business

we analyze the impact of Bluetooth ACL（Asynchronous Connectionless）data packet on the throughput of the system under different Signal to Noise Ratio（SNR）

and propose a selection algorithm based on SNR linear predictive adaptive packet.The algorithm uses the Kalman prediction method to predict the SNR of future time slot based on time window

according to the estimated SNR to determine whether the system is required to switch the packet in the next time slot

and which packet types are more suitable for switching.Therefore

the throughput is optimal.The simulation results show that SNR prediction can respond to the channel changes in advance

and the overall optimal throughput is obtained under different channel quality.