Abstract: Huazhou is an important rice production base in Guangdong Province. Planthopper is one of the major rice pests. The weather condition is a key factor causing the occurrence and development of rice planthopper. The main harm generation of rice planthopper on late rice in Huazhou is the 6^th generation. In order to acquire the rules between meteorological factors and rice planthopper occurrence, the author studied the relationship between the 6^th generation rice planthopper occurrence period and extent and meteorological conditions. Therefore, the author could improve the accuracy of the prediction on rice planthopper occurrence period and extent by using meteorological factors. The author used stepwise regression with SPSS analysis software to analyze the data of the rice planthopper main harm generation’s impact on late rice and corresponding meteorological information of Huazhou in Guangdong Province from 1996 to 2011. Applicable meteorological factors were selected and prediction models were built on the late rice planthopper main harm generation’s adult peak period, nymph peak period, occurrence extent and occurrence area. The author chose 2012 and 2013 data as independent samples to do the model test. The statistical forecast models of adult peak period, nymph peak period, occurrence extent and occurrence area of the 6^th generation rice planthopper were all approved by significant testing at 1% level. Substitute the corresponding meteorological observation data from 1996 to 2011 in the formulas, and the annual variation trend between the measured and simulated values was consistent, and the relative accuracy was 87.5%, 93.8%, 90.9% and 94.2% respectively. The results showed that the measured and simulated values could be in good agreement with the actual values in the year of 2012 and 2013, therefore they could be used in rice planthopper forecast. By stepwise regression analysis of the occurrence and extent of rice planthopper main harm generation (the 6^th generation) prediction examples in Huazhou, the author could find that the prediction ability of this method was good. When there was a high correlation between the selected meteorological factors and the measured values, the author could predict the occurrence extent range. While establishing the model of the rice planthopper occurrence period and extent, the author took the precursor value, therefore the model was more predictable.