The distribution and routing of products within a supply chain has brought with it new challenges such as reducing the environmental impact of its operations and meeting its service levels. These challenges are in addition to the traditional objective of the inventory routing problem (IRP) of reducing the logistics costs associated with the production, maintenance and distribution of products. In particular, IRP problems for perishable products are considered of great importance worldwide, due to the challenges they imply in decision making in a supply chain, especially at the replenishment and production level (Muriana, 2016) and because of their application in different areas such as the distribution and handling of blood, the distribution of radioactive or chemical materials, foods such as dairy products, fruits and vegetables and even in fashion clothing (Coelho & Laporte, 2014)that affect in one way or another the quality of life of people and where consumerism has grown rapidly (Lmariouh et al. , 2017). This paper presents an approach to the IRP problem of perishable products, considering routing decisions associated with the type of fleet to be used in each period, the balancing of its capacity and the determination of attention between nodes, proposing a simheuristic solution method for a multi-objective problem, which considers the reduction of environmental impact, the proportion of backorders against the demand served and the reduction of logistics costs. The problem also takes into account lateral transshipment between links in the supply chain, stochastic demand and waiting times between the different links. The proposed method was implemented in instances of different sizes and its results were compared with the solutions of the mathematical model in its deterministic version as well as with simulations on it and the best solution obtained from the ALNS metaheuristic, obtaining a better performance of the simheuristic for the objective function related to carbon dioxide emissions by about 50% in the medium-sized treatments and 47% in the function related to costs. However, the simheuristic did not perform well in the large treatments and in the objective function related to backorders, which offers an opportunity to generate further research and work in this field.