Background: Dysarthria is a common symptom following stroke and represents an important cause of functional impairment in stroke patients. A better characterization of dysarthria could facilitate differential diagnosis and optimize healthcare service distribution. Aim: To determine the speech characteristics, dysarthria type and severity in the acute phase following ischemic stroke in a comprehensive stroke centre. Methods & Procedures: First-ever ischemic stroke patients consecutively admitted to the Stroke Unit of Ghent University Hospital were included in this prospective clinical study between March 2018 and October 2019. All participants admitted to the Stroke Unit were screened for dysarthria by a speech-language pathologist within 72 h after admission. When dysarthria was identified, speech characteristics were evaluated via auditory-perceptual assessment and objective measurement of acoustic parameters. Dysarthria type was determined based on the Mayo Classification System. Severity of impairment was scored at function and activity level using the Radboud Dysarthria Assessment and the evaluation of speech intelligibility at sentence level using the Dutch Speech Intelligibility Assessment. In addition, dysarthria recovery was separately evaluated in all participants using the National Institutes of Health Stroke Scale (NIHSS) at hospital admission, day 3 ± 2 and day 7. Outcomes & Results: A total of 67 out of 151 participants (44%; mean age = 69 years; SD = 13; 28 females) were diagnosed with dysarthria in the acute phase following stroke. Standardized assessments were possible in 72% (48/67) of participants. Imprecise articulation of consonants, harsh voice quality and audible inspiration were the most frequent observed speech characteristics. The acoustic parameters maximum phonation time and maximum loudness deviated most from normative values. Unilateral upper motor neuron (UUMN) was the main dysarthria type present in 52% (25/48) of participants. A total of 58% (28/48) and 71% (34/48) of participants had no/minimal/mild difficulties at the functional and activity levels, respectively. Speech intelligibility was mildly impaired (median = 91%; IQR = 73-97). According to the NIHSS, sub-item speech score at hospital admission, 46% (70/151) of participants had dysarthria, of which half recovered completely from their dysarthria within 1 week after stroke symptom onset. Conclusions & Implications: UUMN was the dominant dysarthria type, and the majority of participants had a mild dysarthria. Half the participants showed complete recovery within 1 week following symptom onset. The observed speech characteristics mainly reflect impairments in the subsystem's articulation, phonation and respiration. Objective measurements of acoustic parameters corroborate these findings. Future research should focus on longitudinal assessment to investigate recovery of symptoms and the long-term impact of dysarthria on social participation.