In production engineering, monitoring of the grinding process is critical for acquiring information on material removal, wheel performance and workpiece quality. Here, a general model of the power signal and material removal rate is proposed to monitor the internal plunge grinding of a bearing outer race way product. Three continuous grinding cycles after dressing were used to analyse the roughing, semi-finishing, finishing and spark-out process under the same parameters. Based on the actual grinding process, a practical analysis method is applied to improve the general model to more accurately predict the power curve. Finally, estimations of grinding wheel performance and grind quality using the grinding power signal model coefficients are also presented. The experimental results showed that the improved power signal model is capable of solving the industrial problem of multi-stage infeed grinding cycles and improving grind quality.