Abstract: An algorithm for detecting low-altitude wind shear and convergence line with Doppler radial velocities is developed. The algorithm includes two parts:data pretreatment and shear calculating. Pretreatment is in order to smooth some noise in records, make them consecutive and keep the med-scale information meanwhile. Median filtering firstly and then moving average is performed for pretreatment in the algorithm. According to meteorology, convergence is defined radial grads of radial velocity and shear is defined azimuthal grads. Least square method is adopted to calculate shear. Besides of that, the algorithm gives vertical shear also.The number of treating data represents the space related to treatment. Therefore, the effect of number on results of pretreatment and shear calculating is separately analyzed. Both in median filter and moving average, the smoothing effect is more obviously while more data are used. But there are still some differences between two methods. Although median filter do better in smoothing noise and keep med-scale trend of data, the sequent is that data become discontinuous. With median filtering means, the best result is btained that the meso-scale movement of data is slippery, consecutive and preserving. Just as pretreatment, different number will lead different outcome of shear calculating. The value of shear will become large while the amount of data calculated is small. And the shear line or convergence line will become narrow. Furthermore, a gate value of result show must be set to help forecaster recognize the line easily, which is detected by the algorithm. In the end, only when the amounts of both pretreatment and shear calculating are all suit with scale of the line which will be detected, the production will be perfect. After theoretical discussion, the algorithm is tested on some heavy rain practically. The first one is caused by a squall line, the second is about of a strong convective line, last is a precipitation of mixing cloud. In all of these cases, the algorithm woks very well. First, both magnitude and position of wind shear or convergence line which are detected are all consistent with the actual. Secondly, high shear value of shear or convergence line is corresponding to reflectivity of heavy rain. Thirdly, Different gate is necessary for different rain. Fourthly, vertical shear exhibits the variety of velocity fields in different elevation. As a conclusion, it is feasible to detect low-altitude wind shear or convective line using Doppler radial velocities. The production will provide an important and objective evidence for weather forecasting and alarming.