Homegardens are defined as less complex agroforests which look like and function as natural forest ecosystems but are integrated into agricultural management systems located around houses. Study on the factor affecting the diversity of plant resource in homegardens is paramount important to improve productivity and sustainability. Previous studies related to homegradens analysis are conducted using ordination techniques (e.g. Principal Component Analysis, Correspondence Analysis). In this study, we introduced the application of Self-Organizing Map (SOM) a type of Artificial neural networks (ANNs) to analyze the effects of socioeconomic variable and homegardens characteristic toward diversity of plant resource and to investigate the spatial configuration occurring within homegardens. The inter-relationships among the socioeconomic variable and homegardens characteristic were extracted and interpreted using the pattern analysis visualized in component planes. Sequential agglomerative hierarchical non-overlapping (SAHN) clustering technique was also used to verify results obtained from SOM by using the unweighted pair grouping method with arithmetic-mean (UPGMA). Ten homegardens were identified from SOM U-Matrix and each of the homegardens was investigated for their horizontal and vertical profile. Inspection of SOM indicates that the region with high d-values for size of homegardens coincides with those of food, ornamental, and medicinal plants. Region of high d-value in Shannon index coincides with region of high d-value in Evenness index. Region of low d-value in income coincides with high d-values in both the Shannon and Evenness indices. Region of high d-value in age of household also coincides with high d-values in both the Shannon and Evenness indices. Combination of SOM, SAHN and spatial analysis techniques has a potential to analyst and monitor not only the factors affecting homegardens biodiversity but also their development and improvement which to our best knowledge has yet been reported in literature.