To make news communication better adapt to the era of big data, it is necessary to innovate news communication. Accordingly, this paper constructs a big data analysis platform for news communication innovation methods based on the regression analysis model in big data analysis, determines the parameters and coefficients of the regression model with the help of the maximum likelihood estimation method, and determines the form of the regression model. Based on the platform, the main paths of news communication innovation are analyzed, standardized regression coefficients are calculated, and the main methods of news communication innovation are changing news communication forms and news communication media. To verify that the above methods can make news communication innovative, we designed a visualization information platform to change the news communication format and launched a simulation experiment to analyze the effect of news communication under different media. The results showed that after the visualization of the news communication format, the number of viewers of news programs increased from 2.45 million to 3.65 million, an increase of 1.2 million. The conversion rate of audience behavior of news programs increased from the original 63% to 87%, an increase of 24%. From the above results, it is clear that news communication under the big data approach has improved from a theoretical definition to an operational route.