Background: Ultrasonography (US) is a widely used diagnostic tool in physical therapy. One of the US variables often utilized to quantify the dimensions of the subacromial space and its association with shoulder pathology is the coracohumeral distance (CHD), however, this measurement presents diverse evidence in terms of reliability. Objectives: To assess the intra- and inter-rater reliability of both expert and non-expert raters when measuring CHD through US in asymptomatic subjects. Additionally, we determined the effect of rater experience and measurement conditions on the recording of CHD. Methods: CHD of 15 individuals were recorded from US images of the glenohumeral joint of both upper extremities in three different positions. An expert and a non-expert rater in US usage recorded three CHD measurements, after a randomization procedure. To determine intra- and inter-rater reliability, the interclass correlation coefficient (ICC) and a multivariate variance model for the effects of rater experience, joint position and time of measure were used. Standard Error of Measurement and Minimal Detectable Change was also estimated for CHD measurements. Results: Intra-rater reliability ranged 0.970 to 0.998) and Inter-rater reliability ranged 0.48 to 0.876). Joint position (F2;55 = 38.308; p < 0.001; eta 2p = 0.582) and measurement time (F2;55 = 6.019; p = 0.004; eta 2p = 0.180) effect was observed on CHD. Conclusion: Excellent intra- and poor to moderate inter-rater reliability between expert and non-expert clinicians was determined, the latter being influenced by the position of the glenohumeral joint position at the time of US recording and the time of recording the measurement.