Based on 30 countries from 2001 to 2013, this paper measures the difference and spatial correlation in pension through global Moran 's I index and analyzes the factors and their decomposition in the multidimensional distance by the regression model and the B-N model. The empirical results show that: firstly, pension in the worldwide has significant difference, at the same time and still has a positive and significant spatial correlation, further B-N decomposition showing that the spatial correlation is more from the trend component; secondly, it has a significant positive effect on its own space, especially the B-N decomposition further shows that this effect is not only from the trend component, but also from periodic components; thirdly, the unemployment rate, expectations of inflation and the government's stability can significantly prevent the development of pension, mainly through the trend component, especially random component to achieve this, good to dig out the root cause of the differences in the development of pension. Finally, this paper puts forward the countermeasures and suggestions for the development of pension in China in order to promote our country to build a more perfect security system for the aged.