This study explored the effect of three different shopping modes, namely shopping alone, shopping with a companion, and shopping using a shopping aid device, on the efficiency of in-store shopping and compared the behavioral difference between these three shopping modes using in-depth interviews with visually impaired people, their companions, and sales agents. The goal is to understand the current condition of shopping of visually impaired people and their related needs. Lastly, the researchers of this study designed and developed the prototype of a smart shopping aid wearable device for product recognition for visually impaired people. The study participants were six students, four males and two females, from Huei-Ming School and Home for Blind Children. For all participants, their task performance accuracy rate and task performance time were recorded and then analyzed by one-way repeated measure ANOVA to determine if shopping time was affected by the mode of shopping. System Usability Scale (SUS) was applied to determine the usability score of the wearable device of this study. The study results are as follows: (1) When comparing the average number of correct task performance between the three modes, the difference was statistically significant. From the post-hoc comparison, the correct rate of using a wearable device for shopping (95%) was significantly higher than of shopping alone (75%) and shopping with a companion (75%). Moreover, there was no difference between shopping alone and shopping with a companion. (2) For the task performance time, no significant difference was found between the variables. (3) The average SUS score was 74.2 (grade C) and the percentile rank was 71. This finding indicates that the wearable device developed in this study is easy to understand and easy to use. The participants showed a short learning curve and a high use intention.