Even with correct training, up to 25% of children never master handwriting like their peers. While research shows a correlation between handwriting difficulties and school failure, these difficulties can also impact children in their self esteem and behavioral development. Since the mastery of handwriting requires a lot of different skills, it is never easy to understand where a given child is facing difficulties, nor how exactly to help him/her overcome them. For this reason, it is of prime importance to detect and understand children’s handwriting difficulties the earliest possible in order to propose the most effective remediation possible. In this thesis, we first introduce a modernize version of the currently adopted handwriting tests, that show clear limitations in the era of digitalization. Indeed, the nature itself of these tests, conducted on paper, restricts them to the analysis of the final static aspect of handwriting. Its dynamics, found to be very important, is therefore hidden and cannot be taken into consideration. For this reason, we designed in collaboration with therapists several features that describe different aspects of handwriting, which are not limited to static but also capture kinematic, pressure and tilt. The designed features have the main advantage to describe very low level aspects of handwriting, which makes them quite independent of the writing content. We verified this hypothesis by giving the proof of concept that our model for automatic detection of handwriting difficulties can be translated from the latin to the the cyrillic alphabet. In the same way, we demonstrated that our model can also, given retraining, be used on paper or directly on digital tablets, like iPads. Finally, we introduced our iPad-based test allowing to extract the multidimensional handwriting profile of the child. This test aims to answer the first of the afore-discussed problems, by allowing to extract the specific strengths and weaknesses of a child, in less than a minute, on different aspects and at different granularities. The second part of this thesis tackles the problem of designing remediation activities for handwriting difficulties. We designed activities specifically targeting the handwriting aspects identified by the model and obtained a preliminary proof of concept that serious games targeting specific skills of handwriting (e.g. pressure, kinematic, tilt, ...) can have a positive impact on the overall quality of handwriting. Finally the two last chapters of this thesis tackle two corollary, but still crucial questions related to handwriting remediation. After the integration of these remediation activities in a Child-Robot Interaction scenario, in which the child is the teacher of the robot, we gave the proof of concept of the importance of the design of robot behaviors towards social acceptance with children, something especially important knowing the importance of the child’s perception of the robot and the interaction with it in such a scenario. Finally, in the last Chapter, we investigated whether it is possible to "remediate some handwriting difficulties by preventing them", i.e., by supporting pre-school children in the acquisition of the fundamental visuo-motor coordination skills required by handwriting.