Unconventional reserves forecast is one of the most challenging issues especially in low permeability formations containing Tight Oil. In the literature, there are many methodologies, so in this work, the methodology used is the 3-Segment Decline Curves Analysis (DCA) through a mathematical model using Python language program to improve the precision of reserves estimation in tight oil reservoirs. This methodology was applied to three oil wells (Y-1, Y-2, and Y-3) located in the Peruvian northwest at Talara Basin, and the initial data for the model was the first five production years. In the productive life of these wells are identified three flow regimes: the Primary Transient Flow (Segment 1), the Boundary-Dominated Flow (BDF) (Segment 3), and the transition period between both (Segment 2). The data matching results for well Y-1, Y-2, and Y-3 in the first five production years were 89%, 95%, and 86% respectively. The results for 30 years forecast show a EUR relative error of 0%, 3%, and 9% for each well compared with the actual production data. Also, the 30 years forecast was calculated using 1-Segment Arps’ hyperbolic equation for each well, and the relative error was 21%, 15%, and 22% each. Therefore 3-Segment results show less relative error than the 1-Segment method, also prove minimal uncertainty, so we can affirm that it is a reliable method to forecast reserves in formations with similar characteristics.