The electrical signaling system in plants represents the most efficient means for rapidly transmitting information about changes in the environment to all plant parts. Recent studies have shown that the application of machine learning techniques to the electrophysiological signal acquired on tomato plants growing under typical production conditions enables highly accurate detection of stress in plants due to either drought, nutrient deficiency, or pest attack. To better understand how specific are the acquired learnings to tomato plants only, this study aims to explore the extent of the universality of the electrophysiological signal from tomatoes and eggplants. To this end, we modeled the drought response in both tomato and eggplants individually, using recordings from 34 plants from each crop, and evaluated the performance of the classification models trained on data from one crop to the data from the other crop. Different features appear as the most discriminative for each crop. Therefore, several models were taken in this analysis, namely those trained with: i) all extracted features, ii) the most discriminative groups of features for the tomatoes, iii) the most discriminative groups of features for the eggplants, and iv) the union of the most discriminative groups of features for both crops. The obtained findings showed that the models built on data from one crop are able to predict the plant state of the other crop if they are trained with the set of features enclosing the most discriminative ones for the crop on which the model is being evaluated. Such findings imply some similarities in the electrophysiological signals acquired from these two crops with a certain level of crop specificity indicated by the dissimilarities between the discriminatory information for a specific stressor.