Previous studies have shown that the shape effect exists in comparing or recognizing simple black and white line drawing objects in the visual metaphor processing, but not for concrete objects. However, all objects in our world are either natural or artificial, therefore doing research on concrete objects seems to be a great significance in helping us to know the world. In the present study, three experiments were conducted to explore the shape effect in four different experimental conditions: conceptually and perceptually similar, conceptually similar and perceptually dissimilar, conceptually dissimilar and perceptually similar, and conceptually and perceptually dissimilar. In this way, we could explore the role of the perceptual similarity of concrete objects’ shape in visual metaphor processing, especially when common categories were absent. Experiment 1 and 2 demonstrated whether perceptual similarity affects the visual metaphor processing, and we used the production task during the third experiment to explore how perceptual similarity affects visual metaphor processing, and what inner psychological processing is. Based on the masked priming paradigm, we used similarity judgment tasks in the first experiment, matching pictures were presented for a duration of 48 ms only and the participants needed to judge whether the picture pairs could be used for the same purpose. The result showed that no matter the concept was the same or not, similarity in shapes had a “facilitation effect” on individual object recognition, and the functional consistency judgment of shape similarity was significantly faster than that of shape dissimilarity. In the second experiment, the object pairs were simultaneously presented for 2 seconds, during which participants were asked to perform the 9 points Likert scale to rate the functional consistency. The result suggested that in the absence of category, individuals would be experienced uncertainty and had a moderate rating when they made judgments on the function of objects which had similar shapes. A production task was applied in the third experiment, within a time frame of 20 seconds, participants were asked to describe pictures presented with pairs, in order to test whether shape similarity affected the number of correspondences between two objects, then we analyzed the correspondences and speech onset times. The results suggested that when objects were similar in shape, individual speech onset time was shorter than when the shape was dissimilar, and much more consistent descriptions would be produced. When common categories were absent, individuals would try to build an ad hoc category to achieve object classification. In conclusion, according to results of the three experiments, physical shape similarity is more “conducive” for object recognition and visual metaphor processing.