Updating is one of the most important executive aspects of working memory, through which old and irrelevant information is replaced with more important one. Cognitive training is an intervention through which working memory and updating abilities can be supported and enhanced in middle and older adulthood. However, research shows inconsistent results regarding their efficiency, and the aim of this study is to verify the efficiency of updating working memory training. Total of 42 participants divided into experimental (n=22) and active control (n=20) group took part in the study. In a period of 10 weeks (20 sessions), the experimental group trained updating by practicing the affective n-back task, while the control group participated in communication skills training. The efficiency of the training was investigated using three measures of near transfer, which participants have completed before and after the training: a non-adaptive version of the trained task, a n-back letter task and a memory update task. The results show better results of the experimental group in the non-adaptive version of the trained task and n-back letter task, while no improvement was found in memory update task. No changes were found in the results in the control group. The paper discusses possible explanations for the lack of transfer to the memory update task, the main point being the impossibility of using the same strategy that was used in the n-back training task.