Introduction: Painful shoulder on hemiplegic patients has been associated to subscapular muscle spasticity. An alternative for treatment is based on subscapular nerve block using phenol. However, there is lack of information on anatomical references for subscapular nerves blockage technique. The aim of this study was to determine mean values and confidence intervals for maximum penetration points in order to facilitate blockage during anesthetic procedure of subscapular nerves. Material and methods: Using 30 dissected adult cadaver limbs, the medial edge of scapula and a horizontal plan to the lower edge of the bone spine were identified. The maximal and minimal angles of the penetration points of the subscapular nerve both above and below the horizontal plan were measured, as well as the minimal and maximal distances from the medial edge of scapula. Superior and inferior bisector angles and the mean horizontal distance (in mm) were calculated for descriptive analysis. Kolmogorov-Smirnov normality test indicated the normal distribution of the data (p > 0.05). Results: Mean superior angle was 6.63° (confidence interval 4.52-8.75), mean inferior angle was 11.30° (confidence interval 8.73-13.86) and mean horizontal distance was 72.53 mm (confidence interval 69.25-75.80). Conclusions: According to this data, for maximum points blocking after solution injection, the needle should be introduced horizontally at the scapula spine level under its medial edge to a mean depth of 72.53 mm. Then, the needle must be driven upwards in a 6.63° angle and later, driven downwards to form a 11.30° angle with the horizontal plan. Those mean values represent 95% of the distribution.