In order to improve the prediction accuracy of the growth rate of Populus ussuriensis, a method for predicting wood properties based on the combination of improved butterfly optimization algorithm (IBOA) and radial basis function (RBF) neural network was proposed. The population in the standard butterfly algorithm was initialized by using the good point set method, and the artificial butterfly algorithm was further optimized by combining the adaptive switching frequency and Levy flight. A new prediction model of wood properties based on IBOA-RBF neural network was constructed, and the results were compared with those of RBF neural network optimized by other algorithms. The results show that the prediction effect based on IBOA-RBF neural network model is the best, the convergence speed is reduced from 37 steps to 23 steps, the prediction result error reaches 5.72%, and the prediction accuracy is the highest. It is concluded that the improvement of the butterfly algorithm is feasible, and plays a great role in the targeted cultivation of populus ussuriensis by relevant personnel.