To evaluate cognitive function with ease at bedside, we developed a novel neurologic exam called the "food memory test (FMT)" and evaluated its validity for use in clinical practice. In this prospective study in a neurology clinic, we asked patients about what they had eaten for the soup and main dish at the last meal [FMT1] and the second-to-last meal [FMT2]. If they answered correctly for both the soup and main dish, they received a "pass" score. If they did not answer or chose the wrong food, they received a "fail" score. We also performed conventional cognitive function tests for comparison. A total of 27 patients was enrolled, and 12 (44.4%) passed the FMT1 test. FMT1 has a strong correlation with conventional memory function tests, including time-place orientation, three-word recall, the Seoul Verbal Learning Test, and the Rey Complex Figure Test . FMT1 was not correlated with a depression score or with frontal lobe function tests. FMT2 showed less significant correlation with conventional memory tests. These results suggest that FMT1 is a reliable bedside test to evaluate recent memory. Clinical application of FMT in daily clinical practice is warranted.