As a cactus plant, the dragonfruit plant is leafless and mainly relies on succulent stems for physiological functions such as photosynthesis and transpiration. There are obvious differences in tissue structure and morphology between the succulent stems of dragon fruit and the leaves of common green leaves, and there is also obvious differences in plant canopy structure, which will directly affect the spectral characteristics of plant canopy. Furthermore, it affects the monitoring of photosynthetic pigments based on spectral technology. In order to explore the method to improve the estimation accuracy of chlorophyll content in the stem and branch of dragon fruit, taking the planting base of Yanshan dragon fruit in Longping Town, Luodian County, Guizhou Province, as the experimental area, the tissues of the stem branch and branch of dragon fruit were collected and determined by ethanol extraction. The chlorophyll content of the tissue was determined by ethanol extraction. Then the spectral data were processed and analyzed by traditional mathematical transform, continuous wavelet transform, discrete wavelet transform and discrete wavelet-differential transform respectively. The correlation analysis algorithm was used to extract and screen the sensitive feature bands. Finally, the partial least square algorithm is selected to construct the estimation model of chlorophyll content in the stem and branch of dragon fruit. The conclusions were as follows: (1) under the discrete wavelet-differential transform algorithm. The peaks and valleys of high-frequency and low-frequency information appear alternately, and the segments of available information have strong stability. With the increase of scale, the amplitude of the curve increases, and the frequency decreases. (2) the methods of differential transform, continuous wavelet transform, discrete wavelet transform and discrete wavelet-differential transform in mathematical transform can improve the sensitivity of spectrum to chlorophyll content in stem and branch of dragon fruit, among which the method of discrete wavelet-differential transform was the best, and the determination coefficient of the spectrum and chlorophyll content in stem and branch of dragon fruit could reach 0.565 (located at H1 decomposition scale 737.5 nm). (3) discrete wavelet-differential transform can effectively improve the ability of the spectrum to estimate chlorophyll content in stems and branches of dragon fruit, and the estimation model based on the H2 scale of discrete wavelet-differential transform was the optimal model. This study analysed the effects of four kinds of spectral processing algorithms on improving the sensitivity and estimation ability of spectrum to chlorophyll content in stem and branch of dragon fruit. The results show that the discrete wavelet-differential transform algorithm proposed in this paper can effectively improve the ability of the spectrum to estimate chlorophyll content in the stem and branch of dragon fruit, which provides basic technical support for the non-destructive diagnosis of chlorophyll content in stem and branch of dragon fruit.