In the present study analyzes the fractal characteristic of multi-source information of gas-solid two-phase flow. The differential pressure fuctuation signals in a riser were collected by a data acquisition module with a measured region of 15-55 cm above the air distributor. Daubechies second order wavelet and Hilbert-Huang transform were used to decompose the original signals. The R/S analysis was applied on the decomposed signals. Besides, the gas-solid flow process of this region was recorded by a CCD high speed camera. The box-counting dimensions of the images were calculated as image's fractal dimensions. By combining the box-counting dimension algorithm theory and actual flow process, the fractal characteristic of two-source information sets were connected. The results show that the decomposed differential pressure signals based on different decomposition methods have similar fractal characteristics. As the decomposition scales increase, the decomposed signals show single-double-single fractal characteristic in turn, which correspond to micro-scale, meso-scale and macro-scale effects of the gas-solid two-phase flow system, respectively. The energy distribution of meso-scale signals is higher than 90%,which indicates that the fluctuation of the differential pressure mainly reflects the interactions between the gas and solid phases in the meso-scale. The image's box-counting dimension mainly reflects the proportion and distribution of the bubbles in the image. As the superficial gas velocity increases, the mean value of the box-counting dimension also increases, while its corresponding fluctuation range becomes narrower. Besides, the image's box-counting dimension corresponds to the decomposed differential pressure signals in microscale. Both of these can be used to represent the degree of movement and energy consumption of the discrete particles. It is conducive to deepen the understanding of the complicated flow behavior in gas-solid two-phase systems, provided that the fractal characteristic of multi-source information can be correlated.