Journal of Soil Science and Environmental Management 2016; Vol. 7(9): pp. 123-132This study was carried out to assist in the formulation of conservation technologies for landscape sustained productivity in the Usambara Mountains, Tanzania. Conventional soil survey methods were used to develop a base map on which 55 soil profile pits were randomly located on representative landforms and land use mapping units. Soil samples were collected from topsoils for soil carbon analysis using the wet digestion method. Descriptive statistics and linear regression models were used to establish relationships between landforms, land use and soil organic carbon levels. Results showed that carbon levels ranged between 0.55 and 10.8% for bush land and forest plantations in the plain and plateau, respectively. Under cultivation, soil organic carbon (SOC) levels varied between 1.03 and 6.34% for mid-slopes and lower slopes of the plateau respectively. The average soil organic carbon in the vegetable growing valley bottoms was 4.5% while in the forest plantation was 5.5% with minimum and maximum of 0.8 and 10.8% respectively. Linear regression model analysis indicated that factors influencing variability of SOC apart from land use are: slope form, soil pH, electrical conductivity and CECclay. It was concluded that soil organic matter in the study area is mainly determined by elevation, slope form and type of land use and management. Introduction of soil erosion control measures and incorporation of crop residues to areas where soil organic matter has been depleted were recommended for sustainable crop production.