Aiming at the state estimation problem of a class of nonlinear systems under complex cooperative attack, a distributed consensus-based recursive filtering algorithm is proposed. Firstly, the phenomena of denial of service (DoS) and false data injection (FDI) are described as two random Bernoulli sequences, and a complex cooperative attack model including DoS and FDI is established by using a unified framework. Then, a recursive filter with distributed structure is designed based on the consensus theory, the optimal filter gain is calculated, and the necessary and sufficient conditions for the estimation error of the filter to meet the mean square boundedness are derived. Finally, the indoor robot positioning problem is used to verify the effectiveness of the proposed filter algorithm