Automatic sentimental evaluations for the outdoors, whether they are man-made or natural environments, can provide valuable insight for applications like tourism, marketing, and urban planning, because they help to learn how people perceive, appreciate, and engage with the surroundings. Often, these places are portrayed in social media posts accompanied by a written description that may or may not convey the same emotion as the image. In this work, we study the automatic sentimental evaluations of pairs of photos and texts, where the depicted images are from natural environments – landscapes. During the analysis, the sentiments derived from the image and text are evaluated independently. These individual sentiments are then merged to form the overall sentiment associated with the pair. The analysis provides the sentiment (positive, negative, or neutral) associated with the image, the text, the combination of the image and the text, and the disparity between the sentiments represented in the two. Overall, an ensemble of deep-learning models was used for images, and an ensemble of machine-learning models for text and for the combination of images and text, the latter applied if the disparity justifies that ensemble. According to preliminary findings, for the landscape photo-text combination, in our private dataset, we achieved an accuracy of 78.75%.