It is commonly known that most rocks have a time dependent behaviour. Although generally elastoplastic laws can be used to solve many quasi-static problems for geomaterials, in some cases it becomes necessary to take into account a time effect. This work presents results on the mechanical behaviour of an outcrop Miocene porous chalk, named Pietra Leccese, in a general elasto-viscoplastic field. The Pietra Leccese limestone has been of great importance as construction material in the Salento area, since the renaissance and baroque age. Such rock has been already used for other studies concerning well-bore stability, damage mechanics and compaction problems of oil reservoirs. The experimental study was performed in the AGIP laboratories, through a set of triaxial compressive tests carried out on cylindrical specimens following a particular step-wise loading path: axial, fast loading phases succeeded by constant stress level periods at regular time intervals (3-6 days), during which free viscous deformations were monitored in the samples. The main aim was to discern the elastic, plastic and viscous properties of the rock, focusing on transient (primary) creep behaviour. Results have shown that the viscous phenomena, like creep, strongly reduced the fragile rupture domain and the threshold of failure, extending the domain of plasticisation of the rock (Maranini & Brignoli, 1999). Subsequently, a constitutive equation was formulated, based on the general theory proposed by Cristescu (1989), adapting the model to the features of the chosen rock (Cazacu et alii, 1997, Maranini & Brignoli, 1998). The determination of the constitutive functions and parameters were strictly related to the way in which the tests were conducted. The yield function was obtained from the irreversible stress work, which is a work-hardening parameter calculated from the points of creep stabilisation. The viscoplastic potential has been substituted by a strain rate orientation tensor depending on the first two stress invariants only, therefore assuming an isotropic behaviour of the rock. The verification of the model showed good agreement between experimental data and the analytical prediction. However, some differences arose when simulating rock behaviour under different stress paths. This is probably and mainly due to rock anisotropy, a fact that is in contrast with the isotropy assumption on which the analytical model is based. Due also to the fact that the rock behaviour is strongly dependent upon loading conditions, the complexity of the model has been consequently increased. The results indicate that, despite of its simplicity, the model can be adequately used to estimate time dependent behaviour of rocks thus integrating other engineering tools in predicting rock behaviour.