Stem and progenitor cells provide a promising therapeutic strategy for amyotrophic lateral sclerosis (ALS). To comparatively evaluate the therapeutic potentials of human bone marrow-derived mesodermal stromal cells (hMSCs) and umbilical cord blood cells (hUBCs) in ALS, we transplanted hMSCs and hUBCs and their neuroectodermal derivatives (hMSC-NSCs and hUBC-NSCs) into the ALS mouse model over-expressing the G93A mutant of the human SOD1 gene. We used a standardized protocol similar to clinical studies by performing a power calculation to estimate sample size prior to transplantation, matching the treatment groups for gender and hSOD-G93A gene content, and applying a novel method for directly injecting 100,000 cells into the CSF (the cisterna magna). Ten days after transplantation we found many cells within the subarachnoidal space ranging from frontal basal cisterns back to the cisterna magna, but only a few cells around the spinal cord. hMSCs and hMSC-NSCs were also located within the Purkinje cell layer. Intrathecal cell application did not affect survival times of mice compared to controls. Consistently, time of disease onset and first pareses, death weight, and motor neuron count in lumbar spinal cord did not vary between treatment groups. Interestingly, transplantation of hMSCs led to an increase of pre-symptomatic motor performance compared to controls in female animals. The negative outcome of the present study is most likely due to insufficient cell numbers within the affected brain regions (mainly the spinal cord). Further experiments defining the optimal cell dose, time point and route of application and particularly strategies to improve the homing of transplanted cells towards the CNS region of interest are warranted to define the therapeutic potential of mesodermal stem cells for the treatment of ALS.