This paper investigates the role of adoption of agricultural mechanization on the non-farm labor supply behavior of farm households using a longitudinal data set from Bangladesh. The paper uses an agricultural household model to establish the link between the labor-saving technology adoption decision and the non-farm labor supply behavior. To control for potential endogeneity between the farm mechanization and the non-farm labor supply behavior; we use bivariate probit model (BPM), endogenous switching probit model (SPM) and endogenous treatment effects (ETE) model. The results confirm that labor-saving technology adoption raises both the probability of participation in the rural non-farm sector and the labor-supply to the rural non-farm sector. The average treatment effects (ATE) on the probability of participation in the rural non-farm sector are 0.30 in the BPM and 0.21 in the SPM. The results from the ETE model also confirm that the farm households double their labor supply in the rural non-farm sector, given the adoption of labor-saving technology.