With the development of oil and gas fields entering the middle and late stages, The potential demand for production well logging increases. How to quickly and accurately locate the location of downhole leakage and determine the amount of leakage has always attracted much attention. Noise logging uses spectrum analysis technology, which can effectively solve the above problems by analyzing frequency and amplitude information. The improved wavelet threshold method is used to process the noise logging data, which can effectively enhance the flow noise of the reservoir. According to the characteristics of the processed noise logging data, the qualitative analysis of the injection well is carried out. Combined with the influencing factors of the noise signal, The quantitative evaluation and interpretation of the injection profile are carried out by using the four methods of noise intensity method, noise mean method, noise median method and amplitude area method. The injection profile interpretation method of noise logging is applied in X development unit, and the identification of leakage position and channeling is accurate. The four methods are consistent with the response of isotope logging data. Among them, the accuracy of quantitative evaluation and interpretation results of noise mean method reaches 70%, which provides a reliable method for monitoring downhole fluid flow characteristics.