Purpose. To compare a semi-automatic strain analysis of the left ventricle and left atrium with a manual method in speckle-tracking echocardiography.Materials and methods. A strain of left ventricle and left atrium was assessed in 110 patients by two methods: manual (Q-Analysis) and semi-automatic (AutoStrain). The following parameters were evaluated: LV global longitudinal strain (LV GLS), LA longitudinal strain during the reservoir phase (LASr), LA longitudinal strain during the conduit phase (LAScd), and LA longitudinal strain during the contraction phase (LASct).Results. The ROI correction was carried out significantly more often with the semi-automatic method of measuring LV GLS than with manual (40.1% vs. 16.4%, p < 0.05). There were significant differences in LV GLS average values, LASr values, and LAScd values obtained by the semi-automatic and manual methods. LV GLS average values obtained by the semi-automatic method were lower (18.8 ± 2.8% vs. 20.0 ± 3.1%, p < 0.001), and the values of LASr and LAScd obtained by the semi-automatic method were higher (LASr 31.6 ± 9.8% vs. 30.3 ± 10.8%, p = 0.038; LAScd 17.1 ± 7.1% vs. 15.4 ± 6.8%, p < 0.001) than in manual. Semi-automatic method takes more time for LV strain analysis and less time for LA strain analysis than manual method.Conclusion. The semi-automatic method of LV and LA strain evaluation showed higher reproducibility compared with the manual method. With the semi-automatic method, the values of the LV GLS were lower, and the correction of ROI was required more often and took more time than with manual. The semi-automatic method of LA strain evaluation was characterized by higher values in the reservoir and conduit phases and required less time compared to the manual method. The LA longitudinal strain in the reservoir phase showed the highest values of reproducibility compared to other LA strain paremeters.