Assisted gene flow (AGF) can restore fitness in small plant populations. Due to climate change, current fitness patterns could vary in the future ecological scenario, as highly performant lineages can undergo maladaptation under the new climatic contexts. Peripheral populations have been argued to represent a potential source of species adaptation against climate change, but experimental evidence is poor. This paper considers the consequences of within- and between-population mating between a large core population and the southernmost population the rare Dianthus guliae Janka, to evaluate optimal AGF design under current and future conditions. We performed experimental self-pollinations and within- and between-population cross-pollinations to generate seed material and test its adaptive value to aridity. Seed germination, seedling growth and survival were measured under current and expected aridity. Effects of population type, pollination treatment, and stress treatment on fitness components were analysed by generalized linear models. Relative measures of inbreeding depression and heterosis were taken under different stress treatments. Self-pollination reduced fitness for all the considered traits compared to within- and between-population cross-pollination. Under current aridity regime, the core population expressed higher fitness, and a larger magnitude of inbreeding depression. This indicated the core unit is close to its fitness optimum and could allow for restoring the fitness of the small peripheral population. Contrarily, under increased aridity, the fitness of outbred core lineages decreased, suggesting the rise of maladaptation. In this scenario, AGF from the small peripheral population enhanced the fitness of the core unit, whereas AGF from the core population promoted a fitness loss in the peripheral population. Hence, the small peripheral population could improve fitness of large core units versus climate change, while the contrary could be not true. Integrating reciprocal breeding programs and fitness analyses under current and predicted ecological conditions can support optimal AGF design in a long-term perspective.