In consideration of unfavorable integration between tracking speed and measurement precision of conventional phasor measurement methods based on various constant-interval sampling, a ten-state linear Kalman-filter model for tracking power system RMS voltage, frequency, frequency variation rate as well as initial phase angle under the condition of low sampling frequency (4~8 sampling points per cycle) is presented. The algorithm is derived using the second-order Taylor expansion formula, and then the formulas of parameters above are established. According to the signal model, an appropriate time interval is selected to guarantee the response speed of algorithm to signal mutation. Sampling frequency, data window duration, and tracking precision to mutation signals are studied. Simulation results show that the algorithm has higer parameter estimation accuracy at low sampling frequency.