Abstract: Abstract: Oil is easily affected by various factors such as the occurrence of oxidative rancidity to produce aldehydes, ketones, and other substances in storage，which seriously affects the quality of oil. At present, many detection methods were researched on aldehyde, but all these were high in cost, complicated in operation, and time consuming, and there were no reports found of glutaraldehyde detection. In order to rapidly detect the degree of grease freshness, we analyze the oil degree of deterioration by detecting the content of aldehydes. Schiff reagent (i.e. Fuchsine sulfurous acid reagent) is a colourless liquid that can produce a Schiff reaction with the aldehyde function. The produced reaction bond forming the quinonoid structure of purple pigment made the solution color. The fat content of glutaraldehyde is higher, causing a more obvious chromogenic reaction. The paper according to the chromogenic reaction principle between the oil and Schiff’s reagent, researched a determination to the glutaraldehyde content in oil by measuring the oil color. We designed a new type of oil color detection system based on ARM embedded technology. The system was mainly composed of color data acquisition module, data processing module, and data display module. The core part of the control system is S3C2440, the overall operation control system. The color acquisition module can work in the device out of stray light. The light source is provided by the D65 light source. Oil color information of color sensor TCS230 is found from the D65 light that penetrates the transparent glass container for oil after the color of the light. TCS230 color data transfer acquisition for S3C2440 analysis can detect glutaraldehyde content in the samples through the determination of oil color change. The results showed that with the increase of glutaraldehyde content in the sample, the sample R significantly increased after color stimulus values decreased gradually, and soybean oil, peanut oil, and sesame oil showed the same trend. This trend and visual sensory evaluation of oil samples Confidante sequentially consistent color becomes shallow. There is a power correlation between the R stimulus value (red color) and the glutaraldehyde content in RGB color mode. In 0.0050-0.0500 mol/L concentration ranges, soybean oil content of glutaraldehyde and the stimulation of the R-value coefficient is 0.931, peanut oil coefficient is 0.945, and sesame oil coefficient is 0.956. The grease glutaraldehyde mathematical model is set up in soybean oil, peanut oil, and sesame oil. By measuring the R stimulus value in the oil sample by the oil color detection system, we can determine the glutaraldehyde content in soybean oil, peanut oil, and sesame oil. There is a good linear correlation between the system determination results and the actual values. The soybean oil linear equation is y = 1.400x−2.108, R2 = 0.991; the peanut oil linear equation is y = 1.326x − 1.545, R2= 0.958; the sesame oil linear equation is y = 1.546x−0.927, R2 = 0.916. We can improve the accuracy of detection results of the mathematical model through the elimination of deviation standards by correction. The research method can rapidly and accurately determine the glutaraldehyde content in different oils, which establishes a better foundation for the fast determination of oil fresh degree during storage.