Abstract: We present an improved protocol for the in situ measurement of water potential of plants with an L-51 leaf chamber combined with an HR-33T microvoltmeter. The protocol was verified with several plants grown in different environments. To facilitate the diffusion of vapour to the thermocouple junction of the L-51 leaf chamber, a swab wetted with a droplet of distilled water and attached with fine aluminium powder (10–15 μ in particle diameter) was used for removing the cuticle of plant leaves. For more accurate readout of data, a data acquisition and recording device and computer software were used to record and analyse the results. Results with this protocol were in good agreement with those obtained with a Schollander bomb. Details for determining cooling time during the measurement and sensor contamination as well as the method and prototol for cleaning the contaminated sensor are given with examples and illustrations. Error evaluation and measures to minimize those errors caused by abrupt temperature fluctuations, especially for field measurements, are also proposed. With the new protocol, the in situ measurement of plant water potential can yield faster, more accurate and reliable results, because the impact of cell sap from damaged cells on water potential is eliminated.