Mindfulness has emerged as an important contributor to health and well-being, although its accurate assessment represents an ongoing challenge. The 39-item Five Facet Mindfulness Questionnaire (FFMQ) is a widely used measure of trait mindfulness that includes five subscales: observing, describing, acting with awareness, non-judging, and non-reacting to inner experience. While the instrument has been shown to possess generally acceptable psychometric properties, no work so far has been conducted to increase the precision of the instrument and its subscales in discriminating between individual trait levels. The present study used Rasch analysis to assess the psychometric properties of the FFMQ using a sample of 296 participants, with the intention to improve the scale if necessary. The best fit to the Rasch model for all five FFMQ subscales and the total scale was achieved after minor modifications that involved combining locally dependent items into subtests and removing two items (24 and 32) that critically affected the goodness-of-fit indices. These findings support the psychometric properties and internal construct validity of the modified FFMQ, and ordinal-to-interval Rasch conversion tables are included here that can be used to increase the precision of measurement without requiring any modifications of the original FFMQ response format. These findings have implications for a wide range of areas where more accurate assessment of mindfulness and its facets is necessary.