In order to improve the data transmission convergence ability of Wireless Sensor Networks (WSNs), a data transmission and aggregation algorithm based on sparse dense array transmission mechanism is proposed. A new transmission matrix is designed by introducing the kernel generation function to quantify the connectivity and load between cluster head node and sink node for improving the evaluation ability of cluster head node transmission effect. And the eigenvectors are sorted by column and convolution algorithm to reduce the transmission value of cluster head nodes for effectively reducing the load of cluster head nodes. The tree decomposition mode is utilized to search available Hamiltonian circuits, and a convergence stability method based on path decomposition optimization mechanism is constructed. The Hamiltonian addressing mode is adopted to optimize the data link between leaf node and root node one by one so as to enhance the coverage ability of cluster head nodes and improve the stability performance of data transmission process. The simulation results show that, compared with the current common algorithm of a stable clustering for WSN data based on threshold filtering and fuzzy clustering, and the energy-saving and reliable multi-path data transmission in wireless sensor networks for medical applications, the proposed algorithm has higher concentration of message transmission, stronger control ability on transmission link jitter, and higher network bandwidth.