To address the issues of nonlinearity, time-varying characteristics and new information priority in processing complex data, this paper proposes a new information priority adaptive nonlinear grey Bernoulli model. By integrating the Bernoulli equation, the time power term and the new information priority accumulation operator, the proposed model fully utilizes the effective information within sparse sample, thereby enhancing the adaptability of grey prediction models to nonlinear and volatile time series. This model's hyperparameters are solved using the Particle Swarm Optimization algorithm, and the objective function is optimized. Meanwhile, a quantitative method is adopted to select the modeling sequence of the case study. In the prediction of China's high-tech industry output, the new model is compared with five benchmark models and subjected to robustness testing. The results indicate that the new model outperforms the comparison models in terms of prediction performance, and exhibits strong robustness. Furthermore, the future output value of China's high-tech industry is predicted.