AbstractIn order to improve the analysis effect of risk factors for blood transfusion in preterm infants, this paper studies the natural gradient algorithm based on statistical analysis methods and analyzes the data processing effect of the algorithm. Moreover, this paper combines the analysis of risk factors for blood transfusion in preterm infants to build a factor analysis system based on statistical analysis algorithms and intelligent algorithms, and combines back propagation (BP) neural networks to build the statistical intelligence model. In addition, this paper designs experiments to verify the performance of the statistical intelligence model constructed in this paper, combines statistical methods to perform data processing, and analyzes experimental results through control experiments. From the experimental research results, it can be known that the statistical processing method of risk factors proposed in this paper has a certain effect. Simultaneously, through a large amount of data training research and analysis, it is confirmed that the system constructed in this paper has good data processing effects and can be applied to practice.