Ti-6Al-4V alloy is a light alloy widely used in aerospace and other industries, with its excellent mechanical properties, it has more and more good prospects for development. Fatigue fracture is one of the main reasons for the failure of key aerospace structural components. Fatigue failure is obviously uncertain and sudden. Therefore, the evaluation and prediction of fatigue properties of Ti-6Al-4V alloy has become a research hotspot. In this study, fatigue tests involving stress ratio, stress concentration coefficient and temperature factors were carried out on Ti-6Al-4V alloy, and S-N curves of 8 groups of samples under different working conditions were obtained. The influence of related factors on fatigue properties of Ti-6Al-4V alloy was analyzed. Aiming at the problems of time-consuming and costly S-N curve acquisition in traditional fatigue experiments, a hybrid neural network model was proposed. The model was trained by using the relatively low-cycle fatigue test data of the samples, and the data was successfully derived, so that the fatigue properties of Ti-6Al-4V alloy samples under different working conditions could be evaluated quickly and accurately.