The purpose of this study was to investigate intra-rater reliability and determine the validity of electromyography (EMG) measurements to represent muscle activity and ultrasonography (US) to represent muscle thickness during manual muscle testing (MMT) to external abdominal oblique (EO) and lumbar multifidus (MF). Twenty healthy subjects were recruited for this study and asked to perform MMT at differing levels. The subjects' muscle activity using EMG was measured by a ratio to maximum voluntary contraction (MVC) and root mean square (RMS) methods. The subjects' muscle thickness using US was measured by raw muscle thickness and change ratio of thickness to maximum (MVC) or resting condition. In three trials, measurements were performed on each subject by one examiner. The intra-rater reliability of measurements of EMG and US to EO and MF was calculated using intra-class coefficients. The intra-rater reliability of all measurements was excellent (ICC=.75~.98) in EMG and US. The conduct validity was calculated by one-way ANOVA with repeated measurements to compare whether the EMG and US measurements were different between MMT at different levels. There was only a significant difference between all grades at %MVC thickness measurement of US. These results suggest that a %MVC thickness measurement of US was a more sensitive and discriminate in all manual muscle testing grades. This information will be useful for the selection of US measurement and analysis methods in clinics.