<p>Oil production prediction plays a crucial role in the process of oilfield development, deployment, and production planning, especially when it comes to reservoir management, yet its accuracy is often unsatisfying. To address this, we proposed a bio-inspired model for oil production prediction, which combined a new genetic algorithm, known as DNA-GA, with a radical basis function (RBF) neural network. The DNA-GA was designed based on the regulation mechanism of DNA and was applied to optimize the RBF network together with the gradient descent method. The DNA-GA can effectively improve the population’s diversity and prevent premature convergence, resulting in high convergence accuracy and excellent optimization ability. The tests based on Schaffer and Rosenbrock functions showed that the convergence accuracy of DNA-GA was increased by 65.38% and 67.69%, respectively, compared with the existing optimization algorithms such as the Firefly algorithm, and the convergence speed was also increased by 50%. Applied into real-world oil fields, the bio-inspired model greatly outperformed traditional neural network models in both prediction accuracy and convergence speed. Comparative analyses against the traditional RBF neural network and GA-optimized RBF neural network revealed a substantial enhancement in the prediction performance of the bio-inspired model. Specifically, the mean absolute error (MAE) was reduced by 75% and 50%, respectively. Moreover, convergence speed, as measured by average convergence iterations, was improved by 57% compared to the GA-optimized RBF neural network. These results proved that the combination of the RBF neural network and the DNA-GA algorithm could greatly improve the prediction accuracy of oil production.</p>