Purpose. In the process of learning motor functions, it is important to avoid perpetuating the wrong movements. For this purpose, it is recommended to use a combination of various media: speech, demonstration and practical activity. However, in some cases the learning process can be disrupted by difficult conditions (e.g. buzz at a swimming pool), which cause significant disturbance in perception of verbal information. In this situation, the teacher may use visual communication (in the form of gestures). The aim of this study was to determine differences in the accuracy of learning swimming skills (in the experimental and control groups) after implementing visual information (in the form of gestures) in the experimental group. Basic procedures. The test method used was a pedagogical experiment conducted among 86 people (40 in the experimental group and 46 in the control group). The difference between the groups consisted in different ways of correcting errors in movements: in the experimental group, I introduced an independent variable - visual information communicated by gestures, while in the control group errors were eliminated by means of verbal information. Main findings. Analysis of the results shows that the difference in the coefficients of swimming accuracy in the experimental and control groups is statistically significant and in favour of the experimental group. The result was significantly influenced by the precision of hip joint movements - bending and straightening - reached in the experimental group. Slightly less, but also important were differences in the precision of single movements like arm flexion, dorsal flexion in the talocrural joint and extension of the knee. Conclusions. 1. Visual information transmitted using "language of gestures" affects the accuracy of learning swimming motor activities. 2. In teaching of swimming motor activities, visual information is more effective than the information communicated verbally.