We propose an adaptive fused-lasso based coefficient subgroup approach in the decentralized network system. The major goal is to improve the model estimation efficiency by aggregating the neighbors’ information as well as identifying the subgroup membership for each node in the network. In particular, a tree-based L_1 penalty is proposed to save the computation and communication cost. Also, we design a decentralized generalized alternating direction method of multiplier algorithm for solving the objective function in parallel. The theoretical properties are derived to guarantee both the model consistency and the algorithm convergence. Thorough numerical experiments are also conducted to back up our theory, which also show that our approach outperforms in the aspects of the estimation accuracy, computation speed and communication cost.