Based on the detection of edible oil,a Raman spectrum detection system using the hollow fiber with silver plated in the inner surface as the sensor cavity is established for detecting edible and fried oils.A mathematical model is proposed to describe Raman signal intensity.The fiber length is optimized by simulation results and experimental data.After system establishment and optimization,samples of oil are measured by the system,including:fried oils collecting from two food stands and several kinds of edible oils.Ratio of Raman peak intensities at 3005 cm-1 and 2897 cm-1 is proposed to describe the degree of oil un-saturation.The un-saturation collecting experiment for the soybean oil used shows that un-saturation degree is inversely proportional to frying time.A smart and efficient detection system is provided to analyze oil structures and estimate oil un-saturation,where Raman spectrum characteristic peak is clear and sharp.It has shown potentiality in food safety detecting field.