Rodent spatial memory is commonly tested using the water-maze; however, there is a potential confound of stress on learning in this behavioural paradigm. This is particularly relevant when testing spatial memory in models of neurodegeneration, such as the 3xTg mouse model for Alzheimer’s disease. Here, we first confirmed that 3xTgAD mice express fear conditioning and then compared the performance of young and middle-aged mice on short-duration versions of the radial arm water-maze (RAWM) and the minimally stressful T-maze spontaneous alternation task. Our main questions were: (1) does the reliance on stressors in water-maze training mask the true cognitive ability of 3xTgAD mice; and (2) are 3xTgAD mice similarly impaired in water-maze and T-maze protocols. Firstly, male and female 3xTgAD mice displayed intact freezing responses in both contextual and Pavlovian fear conditions. As male 3xTgAD mice displayed relatively enhanced fear responses the remaining tests were performed using only female 3xTgAD and control mice in order to equate for response to stressors. We found that alternation rates after both short and long delays were impaired at both ages in female 3xTgAD mice, indicative of robust spatial working memory deficits. For RAWM, again performance deficits were found in young 3xTgAD mice. As both tasks had similar efficacy at revealing early spatial memory deficits, we suggest that spontaneous behavioural protocols be prioritised over water maze testing in models such the 3xTgAD mouse as the former provide a far less stressful but equally effective alternative.