Vibrotactile skin-reading effectively conveys rich information via vibrotactile patterns, which has gained attention due to recent advancements. However, training to recognize and associate vibrotactile patterns with their meaning is time-consuming and tedious. The conventional training methods use repetitive exposure of the vibrotactile stimuli along with visual and auditory cues of the corresponding symbol. This work proposes a novel visual-based training method to teach users the associations between semantic information and vibrotactile patterns. Our proposed visual explanation training is compared with the conventional training method in a study with 18 participants. Results show that participants achieve a better performance using the new visual explanation training when identifying single English alphabet characters. Moreover, the proposed training also incurred a significantly lower workload (NASA TLX) and was preferred by study participants. The proposed method is thus effective and offers a less stressful form of training users for skin reading.