Wireless sensor network (WSN) is a kind of widely used collaborative network. The critical nodes detection (CND) helps to analyze the vulnerability of the network, which improves the stability of the network and the efficiency of cooperation between nodes. However, most existing works focus on using network structure indicators such as degree centrality and closeness centrality to analyze the importance of WSN nodes, without considering the cost of defending against attacks to nodes in specific attack scenarios, which limits the practical application of existing methods. To address the above issues, we propose a method named MO-CND to improve the accuracy of CND in WSN. As far as we know, this is the first work that jointly considers network structure and differentiated costs of CND in collaborative networks. First, the simulation modeling of WSN collaborative network is realized by the method based on cellular automata in this work. Moreover, the factors affecting node criticality are analyzed by comparing the network performance based on the simulation model under different attack scenarios. Then, the problem of CND is transformed into a multi-objective optimization problem to simultaneously minimize the destructiveness of attacks and the cost of defending against attacks and an improved algorithm based on NSGA-II is proposed to solve the problem. Simulation experiments are conducted to validate the performance of our proposed algorithm outperforms the baselines.