Study areas: We have carried out the study in four sub-districts (i.e. Bhaluka, Fulbaria, Muktagacha and Mymensingh Sadar) of Mymensingh District in Bangladesh. The district covers an area of around 4363 km2 consisting of several small valleys between high forests (GOB 2009b). The temperature ranges from 120C to 330C and the average annual rainfall is about 2174 mm. Bhaluka is located at 24022.5 N, 90022.7 E and the total area is 444.05 km2. Fulbaria is located at 24038 N, 90016 E and covers an area of 402.41 km2. Muktagacha is located at 24045.5 N, 90016 E and its total area is 314.71 km2. Mymensingh Sadar is located at 24045 N, 90025 E and covers an area of 388.45 km2. The soil formation of the district is flood plain, grey piedmont, hill brown and terrace. Jute, mustard seed, sugarcane, tobacco, sesame and paddy are common agricultural crops cultivated in the study areas, while jackfruit, mango, banana and papaya are the most common among horticultural crops. Previous studies on this aspect were carried out in the southeastern part, northeastern part and northern part in Bangladesh where land value and demographic pressure on forests are comparatively less than our study area, which is located quite close to the capital of Bangladesh. Besides, the study areas are located in and around the state forest areas, which is badly degraded and encroached upon. Therefore, studying the homestead forest floral composition and diversity seems relevant to know the current state of the art and future dynamics of homestead biodiversity. Additionally, this study has given a chance to compare the results with previous studies especially with the National Forest and Tree Resources Assessment 2005–2007 Bangladesh, Dhaka, Bangladesh. Sampling and data collection We have carried out the study in the four sub-districts of Mymensingh District in order to understand the plant species composition, plant species diversity, their abundance and patterns in homestead forests. The study was conducted during March 2010 and investigated in randomly selected 120 homesteads from 12 villages (30 from each sub-district) for sampling. The total human population of the Mymensingh District (consisting 12 subdistricts) is 3,957,182 (GOB 2009b), of which the population number is lower than the average (i.e. 329,765 persons) in seven sub-districts, but higher in the remaining five sub-districts. However, in this study, to facilitate the sampling activities, we randomly chose two sub-districts (i.e. Bhaluka and Muktagacha) among the seven subdistricts consisting of human populations lower than the average, while two other sub-districts (i.e. Fulbaria and Mymensingh Sadar) were chosen from the remaining five sub-districts that represent human populations higher than the average. We have used semi-structured questionnaire and conducted interviews (both formal and informal) and Focus Group Discussion (FGD) for data collection. The questionnaire included socio-economic and demographic information related to household. In addition, a separate questionnaire is used for data collection about the cropping patterns that were available in homesteads. Name and number of plant species in all the randomly selected households were collected during the field survey. Herbaceous plants were not taken into account because of their unknown utility and lack of economic importance to the homestead dwellers. We categorized the land allocation status for home gardening into none (no land for homestead forest), very poor (1–400 m2), poor (401–800 m2), medium (801–1200 m2) and rich (above 1200 m2) (adopted after Khan et al. (2007)). Species diversity The diversity index is a well-known method of showing that diversity is not only numbers of different species, but also how well each of these species is represented in different areas. Species diversity is an index that incorporates the number of species in an area and their relative abundance. It is a broader concept than species richness. Species richness is only a measure of the number of species per sample. The more species present in a sample, the ‘richer’ the sample. Consequently, in the measurement of species richness, the number of each individual in each species is unaccounted. The same importance is given to all species whether they are abundant or not. Species diversity, on the other hand, takes into account the relative abundance of a species and not just its occurrence. If a community is composed of few species, or if few species are abundant, the species diversity will be low. High species diversity indicates the presence of a complex ecological community. Various methods exist to make a relationship between species composition and species diversity. To demonstrate the relationship, Simpson’s diversity index, Shannon–Weaver function and their evenness will be measured.