Abstract Problem: Global poultry production needs to rise by 70% by 2050 to meet growing demand for animal protein, but current hatchery techniques in resource-limited areas are hindered by technological shortcomings which reduce hatchability, raise embryonic mortality and make them economically unfeasible because of unreliable electricity supplies and limited environmental control. Research Gap:Although there has been progress in the development of incubation systems that are wireless and connected to the Internet, two major issues remain: (1) hatchability outcomes vary widely reported from 78% to 94.8% without any systematic comparative data on which to base evidence to optimize these systems, especially those powered by LPGs in off-grid settings, and (2) lack of adaptive control architectures capable of meeting the changing physiological needs of developing embryos throughout incubation. Purpose: This study was designed, developed, and empirically tested an adaptive LPG-fuelled smart poultry incubator with predictive developmental modeling, multi-sensor IoT network, thermal energy storage and real-time mobile monitoring. Methods: The three replications of this controlled experimental design were 420 eggs in each case, in an adaptive LPG system (treatment) and in a conventional fixed-setpoint electric incubator (control). Temperature/humidity stability, hatchability, chick quality, energy consumption, system responsiveness, and user acceptance were evaluated using independent t-tests, Mann-Whitney U tests, effect size calculations, and thematic analysis. Findings: The adaptive system had a hatchability of 95.8% which is 8.9 percentage point higher than conventional (86.9%), a temperature stability of ± 0.25°C (vs. ±0.65°C), a humidity stability of ± 2.1%RH (vs. ±5.3%RH), and was able to complete the critical humidity increase during hatching. Operating cost cut by 47.8%, Cost per chick by 53.2% and System responsiveness was 3 times faster. The acceptance by the users was very high, operators felt more confident and less stressed in their operations. Implications: The adaptive control architecture represents a transformative strategy for precision hatchery management under resource constraints enabling smallholder farmers to achieve commercial grade hatchability at lower operational costs, greater energy reliability and quality of chicks, thus contributing to sustainable intensification of poultry production and food security in developing economies. Originality: This study introduces an adaptive control architecture for LPG-fuelled incubation systems, which is the first empirically validated one, specifically designed for this kind of systems, and based on predictive developmental modelling and real-time monitoring through the IoT in an open, reproducible and low-cost architecture.