The fault diagnosis and reconstruction of sensor in aero-engine control system are related to the safety and stability of the engine. In this paper, a deep neural network based on self-attention mechanism is applied to estimate the analytic redundancy of sensors in aero-engine control system. It can estimate the health parameters of engine gas path components at different heights and Mach numbers, and on this basis, the analytic redundancy of sensors under degradation is estimated. A component-level model of turbofan engine was established and simulated. The results show that: The estimation results, the number of network model parameters and the calculation speed of the deep neural network based on self-attention are all better than those of the convolutional neural network. In the experiment of health parameter estimation and sensor analytical redundancy estimation of a turbofan engine, the average relative estimation error is respectively 0.0034% and 0.0105%.