The success of a company is determined by its marketing success, meaning that every company, before distributing its products, first thinks about access to marketing so that it can be distributed smoothly, because if this is not the case, it will certainly face difficulties when the products are marketed/distributed. This study aims to determine whether there is a smooth relationship between product distribution and consumer loyalty at PT. Putra Ali Sentosa (PAS) Pondok Batu, Central Tapanuli Regency.                    From the calculation of the research instrument, the results obtained between the variables X and Y for each item are declared valid and reliable because the total corrected item value is greater than the critical r value (0.30). And the Cronbach Alpha value is above 0.60. From the normality test, all of them show that the distribution of the data does not violate the normality assumption, namely the histogram diagram is in the shape of a bell and the data distribution is close to the diagonal line in the normal curve of the P-P Plot. There is a positive relationship between smooth distribution and consumer loyalty which can be proven by calculating the correlation coefficient with a strong correlation of = 0.764. From the calculation of the equation of the linear regression line, variable X and variable Y are: Y = 11.577 + 0.810 X. This means that the effect of the dependent variable (consumer loyalty) is determined by the independent variable (smoothness of distribution) with a regression coefficient of 0.810. From the calculation of the coefficient of determination, a figure of 58.4% is obtained, and the remaining 41.6% is determined by other variables not discussed in this study, such as product quality, service quality, product price and other variables. From the hypothesis testing carried out and by comparing the calculated t value with the t table value, it is found that the calculated t value is greater than t table (9.102 > 2.001) this means that the hypothesis accepted is the alternative hypothesis (Ha)