Aiming at the problem of insufficient data mining depth of the Remaining Useful Life prediction model of aero-engine at this stage, a prediction method of dual-channel model is proposed. First, a dual channel network structure is constructed: channel one uses time convolutional networks, which enables the network to have a larger receptive field and computing speed through residual structure and hole convolution; channel 2 uses convolutional long short-term memory network to extract multidimensional spatiotemporal features and capture long-term dependencies of data. Then, the multi head attention mechanism is used to reassign weights to the features of the dual channel network. Finally, the dual channel network is used for feature fusion output to achieve prediction of the remaining life of aircraft engines. Experimental validation was conducted using the turbofan engine degradation dataset and compared with other CNN-biLSTM models, multi feature attention models, multi head attention models, and CNN-GRU models mentioned in literature. The results indicate that the proposed model performs better on all three evaluation indicators, providing a new approach for predicting the remaining life of aircraft engines.