Sputum smear microscopy is the cornerstone of diagnosing infectious tuberculosis. The presence of microscopic errors may misclassify or misdiagnose cases as non-cases, or vice versa. Substandard performance will compromise the efforts to detect tuberculosis and complicate measures to control. This study aimed to determine the proficiency of microscopists at three selected government health clinics in Malaysia. A cross-sectional study was conducted in 2009-2010. Three states were selected based on their high sputum positivity rate. All microscopists were enrolled and instructed to stain and grade a set of seven predetermined densities of mycobacilli slides. Two independent raters assessed their readings. A total of 100 microscopists and 700 slides were tested. 88.2% of slides were in agreement, with sensitivity of 83.8%, specificity of 94.3%, positive predictive value of 95.2% and false negativity rate of 18.7%. From the low positive slides, 27.5% were graded as negative. Two-third of microscopists achieved the accepted grading proficiency and 37% were scored as good staining. There was a need to intensify training on smear microscopy if the gateway for diagnosing TB in Malaysia was smear microscopy.