Rice paddy is a significant source of atmospheric methane (CH4), a major global warming source. CH4 emission from paddy fields is greatly influenced by phosphorus (P) management, especially the long-term non-P application on CH4 emission is largely unexplored. In the present study, long-term non-P application (NK) and P application (NPK) treatments of two paddy fields in Suzhou (from 1980) and Yixing (from 2009), Tai Lake region was done. The effect of P application on CH4 emissions and related microorganisms (i.e., methanogens and methanotrophs) from 2019 to 2020 was analyzed. Results revealed that long-term NK treatment didn't alter the seasonal trend of CH4 flux, but significantly promoted CH4 emissions at the tillering stage. The non-P application for >12 years caused the cumulative CH4 emissions of NK treatment in the whole rice season significantly increased by 41.9–221 % in two fields compared to NPK treatment in 2019 and 2020. NK treatment increased the abundance and diversity of methanogens, while reducing the abundance and diversity of methanotrophs. Community composition of soil pmoA gene differed in two experiment sites. Correlation analysis revealed that the CH4 emission was significant and positively related to soil mcrA gene and C/P while negatively related to soil pmoA gene and P. Structure equation model analysis show the low soil available P content was the dominant driving factor for the high CH4 emission under long-term non-P application through its direct impact on soil mcrA and pmoA genes. The increased soil organic acid content was another driver which was positively related to soil mcrA gene and negatively to soil pmoA gene. Our findings demonstrate the important role of soil P in regulating CH4 emissions from paddy fields in the Tai Lake region, China, and suitable P application is necessary for ensuring the yield while reducing CH4 emission.