Wildfires have a great impact on the global ecosystem and human society, so the prediction and prevention of wildfires is necessary. This article uses the MOD14A2 data, the NCEP/NCAR and ERA5 Reanalysis data, the GFEDv4 data and the Scripps O2 data to analyze the correlation between wildfires, meteorological elements and oxygen concentration in the Boreal North America (BONA), the Temperate North America (TENA), the Australia and New Zealand (AUST). The following preliminary conclusions were obtained: 1) From 2001 to 2015, 2002 was the year with the most wildfires, and august was the month with the most wildfires. Besides, Northern Africa, Southern Africa and South America are the main wildfires-affected areas, the total wildfires area from 2001 to 2015 is about 2148 million ha, accounting for nearly 80% of the global wildfires area in these 15 years. 2) Globally, the correlation coefficient between temperature and wildfires area is 0.47, between wind speed and wildfires area is 0.17, between precipitation and wildfires area is -0.41; between relative humidity and wildfires area is -0.19. 3) AS the direct path coefficients of oxygen concentration are nearly 0.38, oxygen can be regarded as a variable independent of meteorological elements. In BONA, from 2001 to 2015, the correlation coefficient between oxygen concentration and wildfires area is 0.61; In TENA, the correlation coefficient is 0.62; In AUST, the correlation coefficient is 0.6. This study illustrates the importance of oxygen concentration for wildfires. So, it is of great significance to the prediction and prevention of global wildfires.