Dairy cow houses are a major contributor to ammonia (NH3) emission in many European countries. To understand and predict NH3 emissions from cubicle dairy cow houses a mechanistic model was developed and a sensitivity analysis was performed to assess the contribution to NH3 emission of each input variable related to a single urine puddle. Results showed that NH3 emission was most sensitive for five puddle-related input variables: pH, depth, initial urea concentration, area and temperature. Unfortunately, cow house data of these variables are scarce due to a lack of proper measurement methods. In this study we focused on a method to assess the urine puddle depth, which can vary between 0.10 mm and 2.00 mm. Our objective was to develop a measurement method for the urine puddle depth capable of assessing this variable on the floor in commercial dairy cow houses with a measurement uncertainty of at least 0.1 mm. In this study we compared two measurement methods being the balance method as golden standard and the ultrasonic method to use in practical dairy cow houses. We measured water puddles in an experimental setup under various conditions. We concluded that the ultrasonic sensor, attached to an X-Y table, can measure puddle depth and can determine depth differences between puddles both with a measurement uncertainty of 0.1 mm. The comparison between the balance and the ultrasonic method gave a mean difference of <0.01 mm (se = 0.006) in puddle depth; a Tukey mean-difference plot showed that the two methods were proportional and that there was no systematic bias.