Abstract: To get a higher accuracy of rainstorm intensity formula, an optimization algorithm (NicheAGA-CGA) was proposed based on niche adaptive genetic algorithm (NicheAGA) and conjugate gradient algorithm (CGA). The rainstorm intensity formula parameters optimization model was constructed by the rule of absolute and relative mean square deviation minimum rule. The niche technology was introduced to improve the variety and the conjugate gradient was combined with the genetic algorithm to improve the local searching. The advantages of global searching in genetic and local searching in conjugate gradient were combined effectively. The rain data of Beijing, Guangzhou and Zhengzhou were used to optimize the rainstorm intensity formula, and NicheAGA was compared with traditional algorithms and standard genetic algorithm. Results show that the optimized results of NicheAGA are better and meet the standard requirements. NicheAGA can get a higher accuracy and provide reference for the study of rainstorm intensity formula.