Abstract: Target detection in wireless sensor network is widely used in many fields, such as military, ecological, medical and security, and it has highly practical research significance. Traditional centralized algorithm relies on fusion nodes so much that the network built is not robust enough and high false alarm rate is caused by its binary decision. Traditional centralized algorithm’s dependence on network coverage will cause “blind holes” of detection alarm in the network. To solve these problems, a distributed target detection algorithm based on credit degree—k-CD algorithm is proposed. k-CD algorithm runs as follows: First, the algorithm adjusts each node’s credit degree using the neighbor automata with all its neighbors’ credit degrees as the input; then, the nodes which have detected the target form a virtual group and make decision fusions using the method of credit degree matching; Finally, the algorithm solves the “blind hole” problem caused by network coverage through triggered mobile nodes. The simulation results show that compared with the majority voting algorithm (MV), k-CD algorithm can increase an average of 35% of detection probability while reducing the false alarm rate by 62% and with different network coverage degrees, network life-cycle can be prolonged by 44% on the average.