OBJECTIVES: Despite the endorsement by World Health Organization to use Xpert MTB/RIF test for diagnosing Tuberculosis (TB) and rifampicin resistance, its use as a rapid diagnostic test is still at infancy stage in Malaysia.  This study aimed to determine the cost-effectiveness of Xpert MTB/RIF for early detection and treatment of TB in addition to current practice.  METHODS: A dynamic transmission model was built to estimate the disease burden and economic burden of TB over a 10-year period.  We calculated the incremental cost-effectiveness ratios (ICERs), comparing the current practice with Xpert scenario where Xpert MTB/RIF was added into diagnostic algorithm.  The ICER was expressed as cost (USD) per disability-adjusted life year (DALY) averted.  Additionally, one-way sensitivity analysis and probabilistic sensitivity analysis were conducted to assess the individual and joint effects of uncertainty across input parameters.  This study was performed from a government perspective.  A 3% discount rate was used for calculating both the costs and DALYs.  RESULTS: Our study suggests that using Xpert MTB/RIF could reduce the notification rate by 7.8% after 10 years. In the base case analysis, the incorporation of Xpert MTB/RIF to the current practice would lead to an ICER of USD12,437/DALY averted.  One-way sensitivity analysis revealed that Xpert scenario is highly sensitive to rapid progression rate, transmission rate and case detection rate.  At the willingness to pay (WTP) threshold of 1 time Malaysian gross domestic product (GDP) per DALY averted (USD10,427.80 in 2014), the probability of Xpert MTB/RIF being cost effective from the government perspective is 98.86%. CONCLUSIONS: With the current evidence available, our model suggests that Xpert MTB/RIF is cost-effective to be implemented in Malaysia if the WTP threshold is set at 1 time GDP per DALY averted.  This provides valuable information for stakeholders to consider in developing new public health policy to eliminate TB from the country.