Aiming at the shortcomings of traditional insulator pollution monitoring method, based on the optical loss of quartz rods, designed fiber optic contamination sensor, measured the relationship between temperature, humidity, percentage of luminous flux decay and equivalent salt deposit density (ESDD). Using the improved support vector machine (SVM) to training model, through a large number of data comparison, the accuracy of the improved algorithm is verified, compared with the traditional model, the fitting degree of data is higher. On this basis, developed a pollution degree online monitoring terminal, it can collection the temperature and humidity and luminous flux data, and send the collected data to the background. Developed the online monitoring system of transmission line insulator pollution degree online monitoring system, to achieve the calculation and analysis of the collected data, get the realtime ESDD and draw the curve. Test results show that, this system achieves online monitoring of insulator pollution degree, provide a reference for the operators.