Abstract: Farmers’ agricultural practices in Vietnam are highly sensitive to weather, climate variability and climate change. The lack of timely and actionable climate-informed agricultural advice leads to significant input and yield losses, which can render investments in farming unprofitable. Development organizations in Vietnam have provided agro-climate services (ACS) to smallholder farmers on a limited scale. They advocate for the government to consider upscaling the provision of ACS, but a large-scale roll-out could strain the government’s financial and human resources. Evaluating the merits of climate services is challenging, because weather and climate risks, as well as the benefits that information services may provide, cannot be derived from robust existing datasets or predicted with certainty. CARE in Vietnam, a non-government organization, has provided ACS in two communes in Dien Bien District since 2015 and they expect to upscale their intervention. In this study, we used a decision analysis approach to develop conceptual models and probabilistic simulations to conduct an ex-ante cost-benefit analysis of four candidate interventions aiming to scale ACS in Dien Bien District, Vietnam. Our analysis was conducted in collaboration with CARE in Vietnam’s project staff, Dien Bien government staff and other experts. Our simulation results indicated a very high chance (98.35–99.81%) of the ACS interventions providing net benefits. With 90% confidence, investments in ACS would return benefits between 1.45 and 16.02 USD per 1 USD invested. Our framework offers a foundation for the design, implementation and evaluation of ACS. The cost-benefit analysis provides support to the government’s potential decision-making process and suggests replacing deterministic with probabilistic approaches when analyzing uncertain and complex decisions in development planning.