This research addresses the issue of unconventional reserves forecast, especially in low permeability formations containing light oil, also known as Tight Oil. The used methodology is the application of Decline Curves Analysis (DCA) with multi segmentation through a mathematical model using Python language program to improve the precision of reserves forecast in tight reservoirs. This methodology was applied to 3 oil wells (A, B, and C) located in Talara Basin in the Peruvian northwest, the first 5 production years were considered as initial data for the model, which were previously treated to minimize uncertainties. The obtained results of historic data matching for well A, well B, and well C were 90.14%, 93.58%, and 93.60% respectively. This shows a minimum degree of uncertainty, so it can be affirmed that it is a reliable method to estimate oil reserves in fields with similar characteristics.