The present disclosure provides a voice synthesis device, a voice synthesis method, a bone conduction helmet and a hearing aid. The voice synthesis device of the present disclosure comprises an aggregation node and a plurality of detection nodes, wherein the detection nodes are used for detecting external voice signals and transmitting the external voice signals to the aggregation node. The aggregation node is used for filtering the detected voice signals and synthesizing the voice signals sent by trusted detection nodes. Thus, the bit error rate of voice signals can be reduced significantly, and the accuracy of voice signals can be improved.