Abstract: This study was carried out in Dodoma urban district with the aim of estimating Cobb- Douglas Stochastic Frontier Production Function for grape production among smallholder farmers; identifying the production factors with significant influence on grape production; estimating values for technical efficiencies, and determining the factors for technical inefficiency. This study involved a random sample of 126 engaged in grape production for at least two years. Data were collected through interviews and analyzed for descriptive statistics using Statistical package for Social Sciences program version 18. Stochastic frontier production function, technical efficiencies and model for technical inefficiency were estimated using FRONTIER program version 4.1. Results indicated that the amount of grapes produced was positively influenced by farm size (p<0.01), household labour (p<0.01), inorganic fertilizer (p<0.05) and pesticide application (p<0.01). The sum of elasticities for production function indicated increasing return to scale and therefore, farmers were operating in irrational zone of production function. In addition, inefficient effect in production function existed at significant rate (γ = 0.728, p< 0.01). Technical efficiency among farmers ranged from 57.0% to 98.0% with a mean of 77.8%, indicating wider possibility for improvement. Increased education level, age, farming experience, and household size, as well as access to extension services were associated with increased technical efficiency. Thus, grape productivity can be increased through increased farm sizes coupled with optimal use of inorganic fertilizers and pesticides.