Agroforestry is an integrative farm management approach in which trees are deliberately integrated with other crops. Agroforestry systems can be effective if appropriate trees are chosen based on particular environmental and economic factors. However, it is crucial to identify suitable trees for agroforestry implementation (AI). The objective of the current study was to recognize the most suitable trees for AI in the agricultural lands of Nazar Kahrizi (NK) rural district of Hashtroud city, located in the northwest of Iran using a multi-dimensional approach. The study area was environmentally evaluated using ArcGIS, which led to the creation of 16 classes with different features. Then, based on the preference of 126 local farmers (from 26 villages of NK), 19 native trees were selected for AI assessment. These trees were evaluated and compared considering seven criteria (i.e., frostbite resistance, salinity resistance, sensitivity to drainage, storm resistance, drought resistance, preventing soil erosion, and economic benefits). Finally, a flexible multi-criteria decision analysis (MCDA) tool (PROMETHEE II) was applied to provide a complete ranking of preferred trees from the best to the worst for each class. The findings showed that the agricultural lands should be allocated for planting elaeagnus (about 79.6%, 27,446 ha), almond (13.5%, 4619 ha), quince (4.6%, 1573 ha), apple (1.8%, 635 ha), and walnuts (0.5%, 176 ha). Measurements showed that AI with the recommended trees in the study area will lead to CO₂ sequestration of about 12.96 Mg yr⁻¹. The approach used in this study provides a valuable resource for decision-making in AI evaluations and, therefore, contributes to preserving the lands from degradation and ensures sustainable AI.

农林业是一种综合农场管理方法，其中树木与其他作物有意结合。如果根据特定的环境和经济因素选择合适的树木，农林复合系统可以发挥作用。然而，确定适合农林业实施（AI）的树木至关重要。当前研究的目的是利用多维方法，在位于伊朗西北部的哈什特鲁德市纳扎尔卡赫里兹（NK）农村地区的农田中识别最适合人工智能的树木。使用 ArcGIS 对研究区域进行了环境评估，从而创建了 16 个具有不同特征的类别。然后，根据126名当地农民（来自朝鲜26个村庄）的偏好，选择了19种乡土树木进行AI评估。根据七个标准（即抗冻性、抗盐性、排水敏感性、抗风暴性、抗旱性、防止水土流失和经济效益）对这些树木进行了评估和比较。最后，应用灵活的多标准决策分析 (MCDA) 工具 (PROMETHEE II) 为每个类别提供从最佳到最差的首选树的完整排名。研究结果显示，农业用地应分配用于种植胡颓子（约 79.6%，27,446 公顷）、杏仁（13.5%，4619 公顷）、木瓜（4.6%，1573 公顷）、苹果（1.8%，635 公顷）和核桃（0.5%，176 公顷）。测量结果表明，在研究区域采用推荐树木进行人工授精将导致约12.96 Mg yr ^-1的CO ₂封存。本研究中使用的方法为人工智能评估决策提供了宝贵的资源，因此有助于保护土地免遭退化并确保可持续的人工智能。