The piezoelectric coefficient of high temperature piezoelectric ceramics, denoted as Bi(Me)O3-PbTiO3, (Me = Fe, Sc, (Mg1/2Ti1/2)) was investigated as a function of temperature by using a custom-designed test frame. Utilizing laser vibrometry, it was possible to assess the piezoelectric coefficient in situ in the range from room temperature to 500°C. The constraints on the sample geometry as they exist in the commonly used resonance/antiresonance technique such as those encountered during poling were circumvented by the use of the converse piezoelectric effect. Comparison with literature data revealed that the current method is a useful alternative for determining the depolarization temperature (Td), defined as the inflection point in a temperature-dependent d33 plot. Measured Td for each poled specimen was compared with that determined by dielectric permittivity as well as temperature-dependent X-ray diffraction data to understand a possible origin of Td. It was also shown that Td matches with the temperature where the dielectric anomaly initiates, and hence Td from the d33 measurement is consistently lower than that from the dielectric permittivity measurement. It is proposed that this discrepancy in the position of Td is due to the fact that the depolarization occurs in two steps.