Obtaining pulmonary perfusion information from chest electrical impedance tomography (EIT) measurements is of great clinical significance. This study aims to explore a new method for extracting pulmonary perfusion signals from chest electrical impedance tomography (EIT) based on variational mode decomposition, which can directly separate pulmonary perfusion-related signals from the original voltage mixing data before EIT reconstruction and perform imaging. EIT data were collected from six healthy volunteers in two states: normal breathing and breath-holding. The EIT data in the normal breathing state were used to extract the pulmonary perfusion signal by VMD. As a result, this method can effectively and accurately extract lung perfusion-related signals from normal breathing data, and the extracted signal lung perfusion images are highly consistent with the lung perfusion images during breath-holding states. It is expected that EIT can directly obtain images of pulmonary blood flow perfusion distribution from normal breathing, laying the foundation for providing more accurate treatment strategy guidance in clinical practice.