The estimation of feedrate in throughfeed centreless grinding is important to assess the productivity of the process by deciding the machine setup configurations, which result in maximum throughput and minimum form errors. In various engineering disciplines, the estimation of object velocity from recorded video has been widely used. This work is aimed to estimate the feedrate of throughfeed centreless grinding process for different setup configurations using video analysis software by tracking the workpiece motion. The new proposed methodology involves feedrate estimation by analysing the recorded video of workpiece motion, comparison with theoretical feedrate for assessing the accuracy and a regression model to predict feedrate for new setup configurations. A mean percentage error of 11.50% was obtained between theoretical and estimated feedrate, which shows the discrepancy between them. A linear regression model was able to fit the estimated feedrate with a standard deviation of 1.00 and an R2 value of 98.41%. The regression model could predict feedrate of configurations well with a predicted R2 value of 96.82%. It was observed that the estimated feedrate deviated from theoretical feedrate due to the factors such as regulating wheel friction characteristics, contact loss and regulating wheel form errors. The method serves as a quick way to analyse the workpiece motion, while the model built using the video analysis data could estimate the feedrate of the process with a good prediction accuracy. The method can also help in deciding the setup configurations for a pre-decided feedrate and for implementing computer vision techniques for in-process feedrate measurement.